Talk:Somalis in the United Kingdom/Archive 4

Somali Education
Middayexpress - sorry, but you appear to have overlooked a rather crucial point about the survey carried out by Africa Education Trust: that is, it was based on the interviews of a *total* of 356 interviews, "refugees and asylum seekers living in Barnet, Enfield, Haringey or Waltham Forest", of which just 23 per cent were Somalis. Therefore the figures you blithely quote as representing the entire Somali community in fact only represents some 83 or 84 individuals - hardly representative of a community which numbers tens of thousands. We must be clear in presenting this in the article: not to do so, but to present the figures as if representative of the entire Somali group in the UK is inescapably WP:OR. Alfietucker (talk) 17:37, 13 September 2013 (UTC)
 * Most surveys include limited samples. That's how they work since obviously not every last immigrant can be sampled. On the other hand, the IIPR paper did not sample any Somalis for their educational background. It just gives two figures but doesn't indicate which surveys if any those numbers were drawn from in the first place. The Africa Education Trust and London Learning and Skills Council paper, by the way, applies to Somali UK immigrants in general:


 * "Large numbers of the Somali community live in Tottenham, Colindale, Edmonton and Edgware. In general the Somali community are highly educated. Approximately half had completed secondary education and a further fifth of those interviewed had completed university. Most Somalis had attended education or training courses in the UK and over half had done a course at FE or HE level. Most Somali women had accessed education or training in the UK. The vast majority of Somalis have English skills of intermediate level or above. Over half of the Somali community have worked in the UK. 20% of Somalis had worked in professional employment in Somalia. Many of these people had worked as engineers or teachers. Only 4% had found professional employment in the UK. Half had worked in semi-skilled or manual jobs in the UK compared to 13% in Somalia."


 * The foregoing is on page 23, in the section aptly titled Summary of each community Middayexpress (talk) 18:46, 13 September 2013 (UTC)


 * "On the other hand, the IIPR paper did not sample any Somalis for their educational background. It just gives two figures but doesn't indicate which surveys if any those numbers were drawn from in the first place". Presuming you mean IPPR, that's not true. The IPPR report in question is their Beyond Black and White. It uses Labour Force Survey data for the period 2000-04. However, the figures that were previously cited do not refer to all Somalis in the UK, rather than just those who had arrived in the previous ten years. The relevant quote is: "Education levels among the new Somali-born immigrants are the lowest of the countries compared, with the highest proportion of people having no qualifications (50.1 per cent) and the lowest proportion of those having a higher qualification (2.8 per cent)". Cordless Larry (talk) 08:50, 1 February 2015 (UTC)


 * The educational levels of Somalis over a decade ago are outdated and do not reflect the current situation. The 2013 IPPR paper on the educational attainment of Somali students gives a 33% GCSE figure, but it concedes that its data is not robust. I've therefore replaced it with a 2014 La Sainte Union School study, which notes the actual latest GCSE figures for 2010-2012. Additionally, the Lambeth Research and Statistics Unit's detailed 2009 study on Somali students indicates that their attainment is rising and is directly related to relative command of English ("the evidence is that once Somali children reach a competent level of English, they forge ahead in their learning and can reach the highest standards" ). Middayexpress (talk) 18:35, 1 February 2015 (UTC)
 * Which 2013 IPPR report is this? The one discussed above is from 2005. Anyway, there's no reason to remove mention of it. If there is more data now available, that can be added to the article - it doesn't have to replace what is already there. Cordless Larry (talk) 20:14, 1 February 2015 (UTC)
 * I've now added the IPPR and other material you deleted back in, while keeping your additions. I have to say though, the La Sainte Union School is just one school. Isn't it WP: UNDUE to be even reporting what this source says, when the topic is Somalis in the UK as a whole? Cordless Larry (talk) 20:25, 1 February 2015 (UTC)
 * Sorry, the data refer to one borough, not just that one school. Still, I think we need to be careful - what applies in Camden might be different elsewhere (or might not, but it would be better to rely on national sources where possible). Cordless Larry (talk) 20:30, 1 February 2015 (UTC)
 * That editorial is not a reliable source . Neither is the data from over a decade ago; that does not reflect the current situation. Also, the IPPR itself concedes in its paper that its data is not robust. The actual scholastic performance of Somali pupils is covered in detail in the Lambeth Research and Statistics Unit's 2009 study on Somali students in the UK. Additionally, the La Sainte Union School paper isn't for that one school, but rather for Camden as a whole. It notes specific GCSE results in keeping with what the Lambeth Research and Statistics Unit indicates. At any rate, per WP:BRD, changes of that magnitude first require discussion then consensus. WikiProject Africa regular AcidSnow already thanked me for my edit, so that's a start. If additional input is needed, I'll ask if it would be alright for me to alert the Somalinet forum members of this discussion. There are several thousand of them, so their expertise in this area should be helpful. Middayexpress (talk) 21:44, 1 February 2015 (UTC)
 * The Economist isn't a reliable source now? That's news to me. You're also replacing much more than just that, including national-level research by educationalists (e.g. ), with data from just two London boroughs, Lambeth and Camden. That's odd if you ask me, but I'm happy to wait for others to give their views. Cordless Larry (talk) 21:53, 1 February 2015 (UTC)
 * Also, data being old doesn't make it not a reliable source. Why can't the historical situation of Somalis be included in the article, if it is clearly identified which time period the data is from? Cordless Larry (talk) 21:54, 1 February 2015 (UTC)
 * Oh, and can you provide a link for the claim that "the IPPR itself concedes in its paper that its data is not robust"? Cordless Larry (talk) 21:57, 1 February 2015 (UTC)
 * I wrote that that specific Economist editorial was unreliable, not the news organization as a whole. This is because it is an editorial (which aren't reliable sources on living persons per WP:BLP), doesn't name its source for the GCSE figures (its numbers may have been drawn from a 2013 IPPR study ), and there are other factual inaccuracies as well . The 2005 IPPR is a decade old and does not reflect the current situation. Additionally, the 2010 Strand paper is unrepresentative because its Somali sample is small (98 pupils), and the coding it uses is different from that used by the Local Authorities that actually tabulated the total number of Somali students in Table 4 (~33,979 students). Similarly, several of Rutter's suggested main factors for underachievement, particularly her claim that UK-born pupils of Somali heritage do not achieve better exam results than Somali-born children, are not borne out by the comparative data. As the Lambeth Research and Statistics Unit notes, higher attainment among the Somali students is in fact directly related to English language proficiency ("Somali pupils' performance at KS2 increases at the stage of proficiency in English increases. Bilingual Somali speakers who were fully fluent in English were more likely to gain level 4+ than pupils who only spoke English" ). Middayexpress (talk) 23:33, 1 February 2015 (UTC)
 * Thanks for your reply. We now seem to be on the same page - or somewhat more, at least. On the editorial, I take your point, but does BLP apply here, per WP:BLPGROUP? On the 2005 IPPR report, I don't see that it being old discounts it. Sure, we shouldn't suggest that it necessarily represents the current situation, but the article isn't just about the current situation, is it? I would have thought there was space for historical material. On the Rutter point, her argument might not apply in the specific case of Lambeth, but that doesn't necessarily mean that it doesn't apply elsewhere. She's not saying that her argument applies to each and every local area. Cordless Larry (talk) 23:42, 1 February 2015 (UTC)
 * No prob. WP:BLP "applies to any living person mentioned in a BLP, whether or not that person is the subject of the article, and to material about living persons in other articles and on other pages, including talk pages[...] such material requires a high degree of sensitivity, and must adhere strictly to all applicable laws in the United States, to this policy, and to Wikipedia's three core content policies". BLPGROUP instead applies to "material about corporations, companies, or other entities regarded as legal persons. That said, how some pupils may or may not have performed a decade ago has no bearing on and should not be used against those of today, who are necessarily not the same students. The Lambeth Research and Statistics Unit also uses a series of case studies at various schools to represent the Somali student community as a whole. As such, it is not comparable with Rutter's dated speculations. Middayexpress (talk) 01:32, 2 February 2015 (UTC)
 * WP:BLPGROUP also states "A harmful statement about a small group or organization comes closer to being a BLP problem than a similar statement about a larger group; and when the group is very small, it may be impossible to draw a distinction between the group and the individuals that make up the group". I wouldn't say that Somalis in the UK are a small group in this sense. Anyway, I've asked for opinions on the BLP noticeboard on that. The Lambeth study might well be representative of Somali pupils in Lambeth or perhaps even London as a whole, but what is to say it is representative of Somali pupils in, say, Cardiff? I think we need to draw on more sources that give a national picture (which is not to say we can't use the Lambeth source too).


 * Also, the IPPR data is not about pupils. It's about the education levels of Somalis in the Labour Force Survey. Cordless Larry (talk) 01:37, 2 February 2015 (UTC)


 * One of the main aims of the Lambeth Research and Statistics Unit study was to "discover factors which contribute to the success of Somali heritage pupils" in the United Kingdom and "to explore strategies used by schools to raise achievement of Somali heritage pupil". It did this "using a case study approach, [where] 7 primary and 2 secondary schools with high number of Somali pupils were selected". The 2005 IPPR data also pertains to over a decade ago, not now. It is therefore unrepresentative vis-a-vis the present per WP:SCHOLARSHIP. BLPGROUP applies to companies, coporations, etc.. For WP:BLP not to apply here, we'd have to be dealing with non-living persons or entities, which obviously isn't the situation ("Editors must take particular care when adding information about living persons to any Wikipedia page[...] People are presumed to be living unless there is reason to believe otherwise"). Middayexpress (talk) 03:46, 2 February 2015 (UTC)
 * I am not sure if I can help as I have just glanced this info. But if something is 10 years old, it obviously has problems in stat terms in the context of this article. I guess you could state the period in which it was accurate and that would fix that problem. It is also true no stats are inclusive of everyone, and sampling error is pretty common. I dont think I can add more than this as I have not monitored the issues heavily enough to offer more.--Inayity (talk) 04:18, 2 February 2015 (UTC)
 * Middayexpress, I'm not suggesting that the IPPR source is representative of the present. I'm suggesting that it be discussed in the past tense, as per Inayity's suggestion above. May I suggest the following wording?


 * "In the mid 2000s, research showed completed education within the Somali community to be low. For instance, in 2005 the Institute for Public Policy Research published analysis of Labour Force Survey data for the period 2000-04, and found that, of Somali-born immigrants who had arrived between 1990 and 2004 (who made up 761 of 812 Somali-born people in the sample), 50.1 per cent had no qualification and 2.8 per cent had higher qualifications". Cordless Larry (talk) 07:55, 2 February 2015 (UTC)


 * Inayity wrote that such outdated data is statistically problematic. It is also contradicted by the contemporaneous Africa Education Trust and London Learning and Skills Council report, which indicates that: "In general the Somali community are highly educated. Approximately half had completed secondary education and a further fifth of those interviewed had completed university. Most Somalis had attended education or training courses in the UK and over half had done a course at FE or HE level. Most Somali women had accessed education or training in the UK." At any rate, WP:SCHOLARSHIP discourages outdated or controversial material ("some scholarly material may be outdated, in competition with alternative theories, or controversial within the relevant field[...] try to cite present scholarly consensus when available, recognizing that this is often absent"). Middayexpress (talk) 17:36, 2 February 2015 (UTC)
 * That report is only making reference to London and is based on a smaller sample size, though. Moreover, there's a difference between scholarly material being outdated (e.g. claiming the earth is flat) and it being about a previous time period. I still haven't seen a good reason why this article can only include material on the present situation of Somalis in the UK, and not give a historical picture. Actually, I did include the Africa Education Trust and London Learning and Skills Council report in my edits alongside the IPPR one, but you removed it. Cordless Larry (talk) 18:32, 2 February 2015 (UTC)
 * The Africa Education Trust and London Learning and Skills Council report's summary of each community on page 23 actually pertains to Somalis in the UK as a whole. The report directly contradicts the other contemporaneous paper's claims, and both are in any event outdated per WP:SCHOLARSHIP. Middayexpress (talk) 18:42, 2 February 2015 (UTC)
 * They'd be outdated if we were using them to describe the current situation. What's wrong with using them to describe the situation a decade ago? Cordless Larry (talk) 19:43, 2 February 2015 (UTC)
 * The problem is that we're dealing with actual people here and their purported qualifications. WP:SCHOLARSHIP also applies to such outdated, conflicting material in general; it's stated right there. Middayexpress (talk) 20:40, 2 February 2015 (UTC)
 * That's not how I understand that policy. As far as I understand, scholarly material is outdated when it is proven to be false, not when it simply refers to a past time period. Also, I'm yet to be convinced that BLP applies here. Anyway, let's wait for outside opinions since we're having difficulty agreeing on this. Cordless Larry (talk) 20:54, 2 February 2015 (UTC)
 * Outdated means its old and doesn't reflect the present situation. A decade worth of new arrivals will do that. Middayexpress (talk) 21:19, 2 February 2015 (UTC)
 * If that were the case, the article shouldn't discuss the history of Somali migration to the UK. As long as it's clear that the statistics are historical, I don't understand the problem. The article needn't only be about "the present situation". Cordless Larry (talk) 21:23, 2 February 2015 (UTC)
 * No, outdated means that it has been superceded by more recent information. Information about 2001 is not automatically outdated and as long as it is the most recent and it is specified that it is from that date it is not problematic to use it. It requires a reliable source to state that the numbers for 2001 are no longer current. We cannot simply dismiss it because we assume them no longer to be current.User:Maunus ·ʍaunus·snunɐw· 21:27, 2 February 2015 (UTC)
 * Thanks Maunus. I would actually go further and say that even if more recent information is available, it is still acceptable to include older data (as long as the more recent data is cited too), because it might be of historical interest. Cordless Larry (talk) 21:30, 2 February 2015 (UTC)
 * Yes, depending on context that could be the case. Especially if what is being described is a demographic progression or similarly.User:Maunus ·ʍaunus·snunɐw· 21:58, 2 February 2015 (UTC)
 * Yes, outdated means it has been superceded by more recent data. The stats weren't even necessarily accurate a decade ago. Another contemporaneous (not contemporary) example: "Anecdotal evidence suggests that the experience of other Muslim groups – particularly recent migrants like the Kosovars, Somalis and Afghans – is often not dissimilar, though Muslims of Indian (and East African) origin tend to have a higher socio-economic profile and higher levels of educational achievement" . Middayexpress (talk) 22:16, 2 February 2015 (UTC)

I don't think statistical evidence can ever be superceded by anecdotal evidence. Though anecdotal evidence can supplement, critique and nuance statistical evidence. In this particular case I don't see exactly the relevance of the passage you quote?User:Maunus ·ʍaunus·snunɐw· 22:21, 2 February 2015 (UTC)
 * No, nor do I, particularly since the quote is about children, whereas the IPPR data is about the education levels of the Somali-born population of the UK as a whole. Cordless Larry (talk) 22:25, 2 February 2015 (UTC)
 * I've just found data on the qualifications of Somali-born working-age adults in London in the 2001 census (table 5 here). It shows that the Somali-born group had the lowest percentage of people with higher level qualifications of all groups. That broadly fits with the picture suggested by the IPPR analysis of the Labour Force Survey. Cordless Larry (talk) 23:11, 2 February 2015 (UTC)
 * It doesn't fit the Africa Education Trust and London Learning and Skills Council report, though; nor is it consistent with that small sample 2010 study (the one that asserted that Somali parents tended to be better educated than the other sampled first generation immigrant parents). The anecdotal evidence, though admittedly not statistical, likewise suggests otherwise. Middayexpress (talk) 23:36, 2 February 2015 (UTC)
 * No, but they had very small sample sizes compared to the number of people completing the census. Cordless Larry (talk) 23:49, 2 February 2015 (UTC)
 * Yeah, and it's also a good decade older than one of the other papers. Middayexpress (talk) 00:03, 3 February 2015 (UTC)
 * Indeed. As stated many times above, I don't want to present it as current. Cordless Larry (talk) 00:08, 3 February 2015 (UTC)
 * That is irrelevant. WP:SCHOLARSHIP clearly discourages outdated material, which has just been defined as material that has been superceded by more recent information. In this instance, superceded by a good decade. Middayexpress (talk) 01:04, 3 February 2015 (UTC)
 * I am sorry but you have not shown that it has been superseded at all, sinc eyou havent presented any comparable data of a newer date.User:Maunus ·ʍaunus·snunɐw· 18:38, 3 February 2015 (UTC)
 * Please see the thread below on the newer 2010 Strand study that Larry first linked to. The contradictory material is in it on page 141. Middayexpress (talk) 19:08, 3 February 2015 (UTC)
 * Please also see my comments on why that data is not comparable to the IPPR data. Cordless Larry (talk) 20:13, 3 February 2015 (UTC)
 * And please see my note below on the contemporaneous study exclusively on these qualifications, which is indeed not comparable with the IPPR's interpreted data . Middayexpress (talk) 16:38, 4 February 2015 (UTC)
 * If that were what it meant, we would have to delete the data in the table on asylum applications for the years before 2013, because it's been superseded by more recent information. We don't do that, because it's of interest to understand the historical situation. Cordless Larry (talk) 07:30, 3 February 2015 (UTC)

Third opinion. This dispute isn't a great candidate for a third opinion, which is meant for disagreements where only two editors have participated. But since I've already read through everything, I'll give my two cents. Relying on 2001 census data isn't ideal. But that doesn't mean we should uncritically accept the 2002 study, which is just as dated and only studied Somali migrants in four boroughs in North London. If we want to mention it, we must (1) mention its year as we would do for any data, (2) describe its geographic limitations, (3) avoid it dominating the section like it does now, and (4) present it as subordinate, not equivalent, to the 2001 census data.—Neil P. Quinn (talk) 16:23, 3 February 2015 (UTC)
 * Sorry if I asked for a third opinion inappropriately. Actually, User:Maunus contributed after I requested a third opinion, but I see that there were actually a couple of other contributors before then (although the vast majority of the discussion has been between me and Middayexpress). Cordless Larry (talk) 16:32, 3 February 2015 (UTC)
 * Any thoughts on the 2005 IPPR study, which is based on the Labour Force Survey? That's the biggest sample-size study we have for the whole country, as the census data is only for London. Also, any thoughts on use of the Economist article? Cordless Larry (talk) 16:34, 3 February 2015 (UTC)
 * Cordless Larry, no worries! If I had noticed that Maunus posted after the request was made, I wouldn't have even mentioned it. The IPPR report seems like a very strong source, and the Economist article also seems useful. Both seem stronger than the 2002 study, although it may still make sense to include a few points from that one in the proper context—I can't access the full text so I can't say for sure.—Neil P. Quinn (talk) 16:53, 3 February 2015 (UTC)
 * Yeah, the confusion is my fault, I actually came here because I saw it ad WP:3O I just didnt remove it from the listing since I thought more participation would be better still.User:Maunus ·ʍaunus·snunɐw· 18:40, 3 February 2015 (UTC)
 * Neil, the 2005 IPPR report is contradicted by a newer 2010 study that Larry himself linked to. The Economist editorial he links to above likewise does not identify from where it culled its purported nationwide education figure. By contrast, the governmental Lambeth Research and Statistics Unit indicates that no such nationwide statistics are in fact available, nor is the size of the student population indeed even known . Middayexpress (talk) 16:57, 3 February 2015 (UTC)
 * Can I check what source the 2010 one is, Middayexpress? There's quite a few sources being discussed here and I can't remember which one that is just by the year. Cordless Larry (talk) 17:00, 3 February 2015 (UTC)
 * It's the small sample Strand study I mentioned above and to which you responded that the census, though a good decade older, had way more respondents. Middayexpress (talk) 17:10, 3 February 2015 (UTC)
 * Thanks. The issue there is that those two sources are about different things. The 2005 IPPR analysis of the LFS presents data on the proportion of Somalis in the UK who have each level of qualification, from none up to higher degree level, as does the London census source. The 2010 Strand study is about the GSCE results of Somali pupils in British schools. Cordless Larry (talk) 17:14, 3 February 2015 (UTC)
 * The actual passage from the Strand study that I'm alluding to is the one that indicates that the sampled Somali parents tended to be better educated than the other sampled first-generation parents. That, like the 2002 Africa Education Trust and London Learning and Skills Council report, indeed directly contradicts the 2005 IPPR paper's claims. Middayexpress (talk) 17:27, 3 February 2015 (UTC)
 * That's not necessarily a contradiction. Just because Somali parents are better educated than some other groups, doesn't mean that Somalis as a whole have a high level of education. Firstly, not all Somalis are parents, and secondly it might be the case that the groups they're being compared to also have low levels of qualifications. If you want to provide a page number, I could look at the passage in more detail? Cordless Larry (talk) 17:32, 3 February 2015 (UTC)
 * It's on page 141. And yes, it obviously is a contradiction because the 2005 IPPR paper claimed they had the lowest education levels among immigrants. Middayexpress (talk) 17:45, 3 February 2015 (UTC)
 * The IPPR study says that Somalis had the lowest level of qualifications amongst immigrants who had arrived in the past 10 years, not amongst all immigrants, so it's not a contradiction. The lower-qualified amongst the other groups may well have been in the UK longer. Cordless Larry (talk) 17:54, 3 February 2015 (UTC)
 * The 2005 IPPR paper claims that the education levels among the new Somali-born immigrants were the lowest of the countries compared. That is indeed a direct contradiction since the 2010 paper indicates that the sampled Somali parents tended to be better educated than the other sampled first-generation parents, including both the recent and more established immigrants. Middayexpress (talk) 18:12, 3 February 2015 (UTC)
 * Yes, but the Somalis who were surveyed in the 2010 paper aren't all newly arrived according to the definition that the IPPR study employs. The IPPR study indicates that longer established Somali immigrants have higher levels of qualifications. Also, not all Somalis are parents, and the 2010 paper is discussing parents only on p. 141. Cordless Larry (talk) 18:15, 3 February 2015 (UTC)
 * What definition does the 2005 IPPR study employ for its "newly arrived"? It appears to allude to new Somali-born immigrants and adult qualifications like the 2010 report. Middayexpress (talk) 18:31, 3 February 2015 (UTC)
 * It's people who arrived from Somalia between 1990 and 2004. I've just realised that the 2010 source isn't actually talking about Somali parents, but Somali parents of current school pupils. That's an even smaller subset, which further explains the discrepancy with the IPPR figures. Cordless Larry (talk) 19:33, 3 February 2015 (UTC)

That definition of "newly arrived" doesn't sound much different from that of the 2010 report. At any rate, like the contemporaneous Africa Education Trust and London Learning and Skills Council report, a 2002 study by Goldsmith's College similarly indicates high levels of educational qualifications for Somali immigrants during the period. The study exclusively aimed to obtain a profile of the professional capacity of Somali nationals living in the UK. In conjunction with 26 Somali community organizations, the researchers interviewed respondents in seven different cities across Britain, and found that: "59% had a qualification from the Somali region, and 9.5% from outside Britain or the Somali regions. 11.5% had a degree or higher degree from the Somali regions compared to 3.5% from beyond Britain and the Somali regions. Since being in Britain, 14% had gained a degree and 12.5% were studying for a degree." The study also notes that "despite considerable skills and experience brought to Britain through previous employment, 'occupational downgrading' and 'segregation' was prevalent amongst respondents". It seems that this lack of recognition of prior qualifications gained in the Somali region may be why the 2005 IPPR's educational qualification figures are lower. The study concludes that: "the research found that the respondents were highly skilled and educated, often having achieved a large amount of employment experience". Middayexpress (talk) 16:38, 4 February 2015 (UTC)
 * Not that different, no, but it is looking at the whole of the UK unlike the Africa Education Trust report, and includes a much larger set of Somalis than just parents of current schoolchildren, which is what the 2010 report focuses on. The Goldsmiths report sounds interesting. We could include data from that alongside the LFS data from the IPPR report. I'll have a read. Cordless Larry (talk) 17:30, 4 February 2015 (UTC)
 * Unlike the Goldsmiths report (which was in part commissioned by the IOM), the IPPR paper isn't exclusively on Somali qualifications, nor were community organizations involved in its formulation. It's apparently just the authors interpreting other, already published data. Middayexpress (talk) 18:12, 4 February 2015 (UTC)
 * The 2005 report contains IPPR's original analysis of raw Labour Force Survey data, which isn't published as such. It's available to download as a dataset, but then it needs to be analysed using statistical software to produce the kind of analysis that IPPR did. It's worth noting that the LFS is the largest household survey undertaken in the UK, is used to generate official government employment and unemployment data, etc. It's conducted to Eurostat standards and is generally a very highly regarded source of data. Cordless Larry (talk) 21:25, 4 February 2015 (UTC)
 * That is what I just wrote. The IPPR authors published their own original interpretation of other, already published data. They're not simply relaying government figures. Middayexpress (talk) 17:53, 5 February 2015 (UTC)
 * It's not already published in any form that can be understood by a human without the use of statistical analysis tools. The dataset is just a massive file full of coding. In order to produce the kind of statistics that the IPPR did, they have to analyse that primary raw data, which is a skilled job and requires, amongst other things, the analyst to weight the data to make it representative, decide on the definitions that will be used, etc. The government hasn't published the figures that the IPPR have. They made the dataset available, which enabled the calculation of the figures, but they didn't publish the findings on the qualifications of Somalis. This guide will give you some idea of the work involved in calculating these figures. Cordless Larry (talk) 18:55, 5 February 2015 (UTC)
 * To illustrate this with an example, I doubt very much that the British government publishes statistics on the average number of children that Austrian-born people living in Liverpool have. However, by downloading a LFS dataset and doing some analysis using Stata, I could generate those statistics. That doesn't mean that I'm interpreting already published data. I would be analyzing primary data collected by the government but not published in that form. Cordless Larry (talk) 19:04, 5 February 2015 (UTC)
 * But regardless of that side issue, the fact remains that the Labour Force Survey is pretty much the best nationally representative survey that exists in the UK, which is why it is used for so much social scientific research. Analysis produced using it is a very good source. Cordless Larry (talk) 19:06, 5 February 2015 (UTC)
 * Middayexpress, if one of your objections to use of the 2005 IPPR source is that it's old, can I ask why you yourself replaced it in the article with an even older source, from 2002? Cordless Larry (talk) 19:45, 5 February 2015 (UTC)
 * That was of course from before when I was aware of what WP:SCHOLARSHIP actually indicates on scholarly material that is "outdated, in competition with alternative theories, or controversial within the relevant field". Hence, why I later replaced both with newer data. At any rate, the IPPR is not relaying government figures; it is indeed producing its own analysis based on publicly available raw data. You could probably do the same with that raw data, but that wouldn't make either of your interpretations necessarily accurate. By contrast, the Goldsmith researchers directly profiled Somali respondents, so there was no interpretation required on their part. Additionally, their study is exclusively on the professional qualifications of Somali nationals; that was its sole purpose. It was also facilitated by many Somali community organizations, and in part commissioned by the International Organization for Migration . Middayexpress (talk) 19:56, 5 February 2015 (UTC)
 * OK, I think our understanding of how the IPPR produced the analysis is now on the same page. Sure, they could have conducted the analysis inaccurately, but is there any evidence that they did? Has their methodology been criticised in other, reliable sources, for instance? There are potentials for inaccuracy in the Goldsmiths source too - for example sampling bias (e.g. the fact that community organisations facilitated it might mean that those Somalis with close ties to those organisations were selected to be part of the sample over other Somalis not known to the organisations). At least the LFS is statistically representative. That said, there is no reason why we can't use both the IPPR and Goldsmiths studies. Cordless Larry (talk) 20:04, 5 February 2015 (UTC)
 * The Labour Force Survey didn't actually produce those figures; the IPPR authors did based on their own analysis of the Labour Force Survey's raw data. How broad the Labour Force Survey's sample is thus doesn't make the IPPR's analysis any less of a methodologically obscure interpretation. On the other hand, the Goldsmiths report directly profiled actual Somali respondents. It also notes therein several measures that the researchers put in place specifically to avoid any sampling bias. I don't think either paper should be used, as they are over a decade old and don't reflect the current situation and they are in competition with each other. I just linked to the Goldsmiths paper to show you that even at that time, the IPPR paper's interpreted figures were not necessarily accurate at all. Middayexpress (talk) 20:32, 5 February 2015 (UTC)
 * Yep, that's the same as all research using the LFS. The data is made available to researchers who then analyse it and publish the results. There were many actual Somali respondents in the LFS sample. Can you explain in what way the IPPR analysis is "methodologically obscure"? You seem to be suggesting that there are faults with their analysis, but haven't actually specified what these faults are. Cordless Larry (talk) 20:38, 5 February 2015 (UTC)
 * As for the sources coming to alternative conclusions, I quote WP:VERIFY: "When reliable sources disagree, present what the various sources say, give each side its due weight, and maintain a neutral point of view". Cordless Larry (talk) 21:28, 5 February 2015 (UTC)
 * Well, it turns out that the IPPR itself already explained the discrepancy between its own earlier Labour Force Survey-derived tabulations and the actual, higher educational qualifications of its Somali immigrants. It has to do with the LFS' own questions. From the IPPR:
 * "The LFS includes questions on qualifications based on equivalent National Vocational Qualification (NVQ) levels, but using these variables to assess the qualification levels of immigrants is problematic since, due to difficulties assessing the UK equivalent of overseas qualifications, foreign qualifications tend to be characterised under a general ‘other qualifications’ heading. As Manacorda et al (2006) note, when native-born respondents report holding ‘other qualifications’, this is generally indicative of very low-level qualifications, whereas when immigrants’ qualifications are classified in this group they are often of a much higher level. This results in the qualification levels of immigrants being under-estimated. In order to avoid this problem, we have used the measure preferred by Dustmann et al (2007): the average age at which each of our groups left full-time education."
 * The IPPR itself thus instead now uses, and recommends using, the age when the immigrants completed full-time education as a more accurate indication of their general educational levels. It indicates this on its Table 5.4; for Somalia nationals, that age is 17.5. This, then, would be the actual measure to go by . Middayexpress (talk) 22:16, 5 February 2015 (UTC)
 * Yes, although the IPPR study doesn't report a high proportion of Somalis reporting "other qualifications". Many answered the LFS question with "no qualifications". Cordless Larry (talk) 23:00, 5 February 2015 (UTC)
 * That's right, it doesn't. The IPPR indicates that foreign qualifications tend to be characterised on the LFS under a general 'other qualifications' heading. And it's this 'other qualifications' heading that is subsequently misinterpreted because for native-born residents, holding 'other qualifications" usually means that they hold very lower-level qualifications. This is then wrongly assumed for foreigners as well (especially those with unrecognized degrees), the end result being that they are registered as having no qualifications. Middayexpress (talk) 23:50, 5 February 2015 (UTC)
 * Yes, but my point is that doesn't really prove to be a problem here, because the IPPR analysis isn't reporting many Somalis in that "other" group. Cordless Larry (talk) 00:01, 6 February 2015 (UTC)
 * That's a good find. I'm not sure we can imply that that explains the discrepancy (I think stating that in the article would be OR, because there are other possible explanations), and average age of leaving education is measuring something different from the percentage of a group who has each level of qualifications, but I'd support using that source in the article. Given your previous objection to "outdated" sources and that it's from 2007, would you object to that? Cordless Larry (talk) 22:27, 5 February 2015 (UTC)
 * The above is from the IPPR's 2007 report i.e. published after its 2005 paper. Maunus wrote that "outdated means that it has been superceded by more recent information[...] information about 2001 is not automatically outdated and as long as it is the most recent and it is specified that it is from that date it is not problematic to use it[...] it requires a reliable source to state that the numbers for 2001 are no longer current[...] we cannot simply dismiss it because we assume them no longer to be current." That is precisely what has just been demonstrated. The IPPR itself indicates that using the LFS' National Vocational Qualification levels -- as it previously did in its 2005 paper -- is problematic, as doing so underestimates the immigrants' actual qualifications, which "are often of a much higher level". Consequently, the IPPR has abandoned its own earlier methodology for the age when the immigrants completed full-time education. This would thus be the actual measure to go by. Middayexpress (talk) 22:54, 5 February 2015 (UTC)
 * I don't agree that it supersedes it as such, as it's a different measure (an average versus a percentage breakdown of qualifications). One could leave school education at 18 with no qualifications in theory, I suppose. The "are often of a much higher level" point applies to the "other qualifications" category, but that's not what I was proposing we include - it was the proportion who answered that they had no qualifications. I'm also a bit baffled that you were claiming that 2005 was too old but that 2007 is OK. Nevertheless, I agree that this is worth adding to the article. Cordless Larry (talk) 22:58, 5 February 2015 (UTC)
 * 2007 is old as well, but Maunus wrote that such material is still relevant if it hasn't been superceded by anything. This is apparently why you insisted on that 89.3% figure from 2001 for the number of Muslim adherents. At any rate, the IPPR's 2005 methodology is indeed different from its more recent 2007 methodology, and specifically because it scrapped that earlier methodology due to its inaccuracy for a newer one. It's written right there, so I'm not sure what's difficult to understand. Ergo, the IPPR's age when the immigrants completed full-time education would be the actual measure to go by. Middayexpress (talk) 23:28, 5 February 2015 (UTC)
 * Don't get me wrong; I'm glad you want to use the 2007 figures. I just don't understand why you thought a 2005 source had been superseded days ago, when you've only just found this 2007 source that you say supersedes it. But anyway, we don't need to agree on that I suppose. Do you want to add it to the article? Cordless Larry (talk) 23:35, 5 February 2015 (UTC)
 * Actually, it's not me who indicates that the IPPR's 2007 methodology supercedes its own, earlier 2005 methodology; it's the IPPR itself. Middayexpress (talk) 23:50, 5 February 2015 (UTC)
 * PS: I think I've found some data on GCSE results that you might like. It's from 2012, covers the whole of England (sadly not the whole UK) and is published in a peer-reviewed academic journal. I'm just reading the article now. Cordless Larry (talk) 23:35, 5 February 2015 (UTC)
 * Oh? Kindly link to it. I'd like to have a look at it and see if it's official and consistent with the Tower Hamlets local authority's GCSE scores. Middayexpress (talk) 23:50, 5 February 2015 (UTC)
 * I don't know how you'd assess that "consistency". It's an average for the whole country, so it probably won't be the same as Tower Hamlets. Anyway, here it is. Cordless Larry (talk) 00:01, 6 February 2015 (UTC)
 * Thanks. Who calculated the average and what is it, if I may ask? Middayexpress (talk) 00:11, 6 February 2015 (UTC)
 * Well, the author works for our friends at the Lambeth Research and Statistics Unit, as it happens (I've only just noticed that!). I'm not sure yet whether he has calculated the average himself or whether he's just reporting an average from the government statistics - I'm still reading it. Either way though, the source is pretty much the best there is. As WP:RS says: "When available, academic and peer-reviewed publications, scholarly monographs, and textbooks are usually the most reliable sources". As we've established, there's no complete ethnicity data for GCSE results available, because not all local authorities use the detailed ethnicity codes, but the article takes an alternative approach. It uses home language instead of ethnicity, which is apparently recorded by all schools in England. Of course, we'd have to note that it's based on language, not ethnicity (or country of birth, etc.). Does that sound OK to you? Cordless Larry (talk) 00:24, 6 February 2015 (UTC)
 * Yeah, I just noticed her name; she is indeed with the Lambeth Research and Statistics Unit. The paper's methodology sounds interesting. I'd like if possible to read it first. I think I may be able to access a copy by tomorrow, so we'll discuss it then. Cheers, Middayexpress (talk) 00:35, 6 February 2015 (UTC)
 * His, I think. ;-) Here's another useful source. I should have done more searching on Google Scholar before! Cordless Larry (talk) 00:37, 6 February 2015 (UTC)
 * Right. Middayexpress (talk) 00:41, 6 February 2015 (UTC)
 * I'll try to draft something to add to the article here on the talk page tomorrow, depending on how busy I am, to get feedback before hopefully being able to improve that education section. Cordless Larry (talk) 00:50, 6 February 2015 (UTC)
 * What again was the GCSE in that 2014 LRSU paper? Middayexpress (talk) 01:11, 6 February 2015 (UTC)
 * You mean the journal article? I hope you agree that we should be selecting the source on how reliable it is, not whether we like the figure it provides or not! Cordless Larry (talk) 07:14, 6 February 2015 (UTC)

There are at least two different topics being discussed here: the level of education held by Somalis in the UK (this is what the debate about the 2005 IPPR source relates to) and the performance of Somali pupils in British schools. Now that we're making some progress with the discussion, let's separate these two things out with subheadings. Cordless Larry (talk) 09:49, 6 February 2015 (UTC)

Level of education of Somali population in the UK
There are two IPPR studies based on analysis of the Labour Force Survey that give us data on this. One, from 2005, looks at the proportion of Somali-born people who have each level of qualification (subject to caveats about the difficulties of classifying foreign qualifications) and the other, from 2007, gives an average education leaving age. I think we have everyone's agreement to use the latter one in this article. Middayexpress objects to use of the first one, but perhaps we can park that issue for now. Cordless Larry (talk) 12:08, 6 February 2015 (UTC)
 * Actually, it's not me who indicates that the IPPR's 2007 methodology supercedes its own, earlier 2005 methodology; it's the IPPR itself, as shown above. Middayexpress (talk) 16:50, 6 February 2015 (UTC)
 * Yes, they note some caveats about their previous figures. However, they clearly still think the qualifications breakdown is of value, because they return to using it in subsequent research such as this from 2008. Cordless Larry (talk) 17:04, 6 February 2015 (UTC)
 * Per WP:BURDEN, please post a quote or screenshot from that closed access paper demonstrating this. Middayexpress (talk) 18:38, 6 February 2015 (UTC)
 * I will do, but my copy is at work, so it will have to wait until next week. Cordless Larry (talk) 18:45, 6 February 2015 (UTC)
 * Actually, WP:BURDEN says nothing about me having to provide a quote or screenshot, but that the burden "is satisfied by providing a citation to a reliable source that directly supports the contribution". Anyway, I'm happy to go above and beyond. Cordless Larry (talk) 18:48, 6 February 2015 (UTC)
 * That's good, because that verifiability policy indeed stipulates this ("When there is dispute about whether a piece of text is fully supported by a given source, direct quotes and other relevant details from the source should be provided to other editors as a courtesy"). Please oblige. Middayexpress (talk) 20:31, 6 February 2015 (UTC)
 * Yes, except I haven't actually proposed a piece of text based on that source yet. I will do when I have it to hand though, and will happily provide you with quotes to support it as you see fit. Cordless Larry (talk) 22:39, 6 February 2015 (UTC)
 * That is not what I asked. I asked you to prove your claim above that the IPPR returned to using its abandoned 2005 methology in its subsequent 2008 paper. Per WP:BURDEN, please provide direct quotes or a screenshot to that effect. Otherwise, the claim has no substance. Middayexpress (talk) 16:11, 7 February 2015 (UTC)
 * No problem. Since it's hard to quote a chart (although I suppose I could quote the figures from it), a photo would probably be easiest. Does anyone know what the copyright situation is there? Am I OK to post a photo of the page of a copyrighted publication? Cordless Larry (talk) 16:36, 7 February 2015 (UTC)
 * Oh, I see. It's possibly a copyright problem. No worries; I can acess the paper myself. Middayexpress (talk) 18:49, 7 February 2015 (UTC)

Right, getting back to this, I propose that we include information on the education level of the Somali adult population in the UK based on two sources: average leaving age from this 2007 report and level of qualifications from this 2008 report. Both need to be caveated: the average leaving age with the fact that it's not necessarily an indication of qualification achieved, and the qualification levels with the fact that some foreign qualifications are hard to categorise using the Labour Force Survey categories. Cordless Larry (talk) 18:39, 17 February 2015 (UTC)
 * Actually, the 2007 IPPR report doesn't indicate that the school leaving age is not necessarily an indication of qualification achieved. It asserts the opposite i.e. that this is a more reliable gauge of the educational levels of immigrants. In its 2008 paper, it writes that "Bangladesh-born and Somalia-born populations were most likely to have no qualifications at all and least likely to possess qualifications at Level Three and above (A-level equivalent)." However, as can be seen, this is based on the same National Vocational Qualification levels that it indicates are problematic for immigrants since it is often difficult to ascertain equivalent UK qualifications for degrees that they earned abroad. Something like the following would thus be more accurate:
 * According to the Institute for Public Policy Research, the Labour Force Survey relies on questions on educational qualifications that are predicated on equivalent National Vocational Qualification (NVQ) levels. This is problematic for immigrants since it is often difficult to ascertain equivalent UK qualifications for degrees that they earned abroad. Consequently, alongside some local qualifications, foreign degrees tend to be processed as 'other qualifications'. For native-born respondents this generic variable implies very low-level qualifications, while conversely the 'other qualifications' of immigrants are often of a much higher level. The actual qualification levels of immigrants are thus underestimated using this method. A more accurate assessment of the educational levels of immigrants is the average age at which immigrants left full-time education, or their school leaving age. For immigrants to the UK that were born in Somalia, this average leaving age was 17.5 years for 2005/2006. Middayexpress (talk) 20:48, 17 February 2015 (UTC)

Average leaving age is an imperfect measure of educational achievement. In some countries pupils are required to repeat years if they fail to pass exams. They could stay at school to 18 and never gain a qualification. The question in the labour force survey is also imperfect. BrumEduResearch (talk) 11:52, 22 February 2015 (UTC)
 * The IPPR aren't suggesting that leaving age is more accurate as such, but rather presenting it as an alternative measure (it doesn't actually measure the same thing). They return to using qualification levels in a subsequent report, which is why I'm suggesting that we present both. The latter should of course be caveated with the points you make about about the classification of foreign qualifications. Cordless Larry (talk) 20:56, 17 February 2015 (UTC)
 * Per WP:BRD, kindly refrain from adding controversial material without prior discussion and consensus. That said, here is what the IPPR 2007 paper actually indicates: "The LFS includes questions on qualifications based on equivalent National Vocational Qualification (NVQ) levels, but using these variables to assess the qualification levels of immigrants is problematic since, due to difficulties assessing the UK equivalent of overseas qualifications, foreign qualifications tend to be characterised under a general ‘other qualifications’ heading. As Manacorda et al (2006) note, when native-born respondents report holding ‘other qualifications’, this is generally indicative of very low-level qualifications, whereas when immigrants’ qualifications are classified in this group they are often of a much higher level. This results in the qualification levels of immigrants being under-estimated. In order to avoid this problem, we have used the measure preferred by Dustmann et al (2007): the average age at which each of our groups left full-time education."". Middayexpress (talk) 21:34, 17 February 2015 (UTC)
 * I haven't added this material from the LFS yet. Yes, the 2008 report includes a similar disclaimer, but they don't judge the problem to be so severe that they don't report the data. I'm suggesting that we include it, with the caveat. Actually, the issue is about foreign qualifications being classified as "other". As a starting point, we could report the proportion of Somali-born population holding no qualifications, since this avoids the "other qualifications" problem. Cordless Larry (talk) 21:38, 17 February 2015 (UTC)

I added the following text to the article, but it was reverted as "dubious". I'm not sure why - it was suitably referenced to a reliable source.
 * According to a 2008 report by the Institute for Public Policy Research, most early Somali labour migrants to the UK were literate in English or Arabic, but only a small minority had completed secondary education. When refugee movements started to increase from 1988 onwards, a greater proportion of the refugees had completed secondary education than previous Somalis migrants, and some held university degrees. The report notes that the educational profile of Somali migrants arriving in the UK has changed again since the arrival of this early wave of refugees. Many educational institutions in Somalia have been destroyed by the war and while some have reopened, many young Somali migrants will not have had the chance to attend university and "are likely to have had a very interrupted education or none at all".

Any comments on this and why it can't be added to the article? It provides useful context to the education section, I think. Cordless Larry (talk) 21:43, 17 February 2015 (UTC)

Here again is why: Per WP:BRD, kindly refrain from adding controversial material without prior discussion and consensus. That said, Rutter's assertion that Somalis allegedly had "no qualifications" is indeed dubious. As already shown, like the contemporaneous Africa Education Trust and London Learning and Skills Council report, a 2002 study by Goldsmith's College similarly indicates high levels of educational qualifications for Somali immigrants during the period. The study exclusively aimed to obtain a profile of the professional capacity of Somali nationals living in the UK. In conjunction with 26 Somali community organizations, the researchers interviewed respondents in seven different cities across Britain, and found that:
 * "59% had a qualification from the Somali region, and 9.5% from outside Britain or the Somali regions. 11.5% had a degree or higher degree from the Somali regions compared to 3.5% from beyond Britain and the Somali regions. Since being in Britain, 14% had gained a degree and 12.5% were studying for a degree."

The study also notes that "despite considerable skills and experience brought to Britain through previous employment, 'occupational downgrading' and 'segregation' was prevalent amongst respondents". It concludes that: "the research found that the respondents were highly skilled and educated, often having achieved a large amount of employment experience". Thus, as the 2007 IPPR paper explains, the LFS-derived assessment is indeed underestimating the actual qualification levels of the immigrants. Middayexpress (talk) 22:13, 17 February 2015 (UTC)
 * The paragraph I added from the Rutter et al. source isn't based on the LFS. I don't really see it as controversial. She doesn't claim that Somalis have "no qualifications". She says young Somali migrants "are likely to have had a very interrupted education or none at all". Yes, you've found another study that comes to different conclusions. That doesn't automatically override Rutter's analysis. It's older, for a start. Cordless Larry (talk) 22:23, 17 February 2015 (UTC)
 * That's the same paper that claims that "Bangladesh-born and Somalia-born populations were most likely to have no qualifications at all and least likely to possess qualifications at Level Three and above (A-level equivalent)" (I have access to it). This is indeed derived from the LFS. So everything else on the purported educational levels of Somali immigrants in that paragraph by default revolves around this dubious claim. The info should instead be sourced to the Lambeth local authority paper, which is more factual and official. Middayexpress (talk) 22:56, 17 February 2015 (UTC)
 * Their LFS analysis doesn't differentiate between early and later arrivals. The report draws on other sources of information as well as the LFS - a literature review and focus groups with Somalis. Cordless Larry (talk) 23:00, 17 February 2015 (UTC)
 * As I wrote, I have access to the paper. The only thing it claims with regard to the actual educational levels of Somalis is the phrase above on the A-level equivalent; it's on page 4. Middayexpress (talk) 23:15, 17 February 2015 (UTC)
 * Is it possible that you have access to the executive summary, which is online, but not the whole book, which isn't online? Also, the Lambeth report is about a different issue. It's on the performance of Somali pupils in British schools, whereas the IPPR report is about the education level of working-age Somalis. Cordless Larry (talk) 23:21, 17 February 2015 (UTC)
 * The executive summary summarizes the book, and it clearly uses LFS-derived equivalent qualifications. The actual paragraph above is also on prior educational levels of Somali immigrants, which the comprehensive Lambeth local authority paper in part touches on, but more factually. It is also official []. Middayexpress (talk) 16:28, 18 February 2015 (UTC)
 * If you haven't got access to the whole book, you won't have seen this passage, which I based the above suggested text on: "Although the early Somali labour migrants were usually literate in Arabic or English, few had completed secondary education. The educational profile of the Somali community changed in 1988, as a greater proportion of the first refugees had completed secondary education, with some holding university degrees. Since then, the educational profile of this community has again changed. The war in Somalia has gravely affected all education. In the north - now the Republic of Somaliland - most urban schools, as well as higher education institutions were destroyed in 1988. In southern Somalia, education has been completely destroyed by the fighting. Some schools have now reopened, often funded by non-governmental organisations. However, younger Somalis who have come directly from Somalia will not have attended university and are likely to have had a very interrupted education or none at all". I think I have paraphrased it accurately. Cordless Larry (talk) 16:44, 18 February 2015 (UTC)
 * It's as I thought. The only thing it claims with regard to the actual educational levels of Somalis is the phrase from the executive summary on the A-level equivalent. It is interesting, though, how it describes Somaliland -- which is internationally recognized as an autonomous region of Somalia -- as a "republic". There's thus possibly a political pov as well. Funny how that A-level blurb also completely contradicts the actual study by Goldsmiths on the qualifications of Somali immigrants . Middayexpress (talk) 17:46, 18 February 2015 (UTC)
 * I think you did a good job paraphrasing Rutter, Larry. She is one of the main experts on Somali education in Britain and this would be good context for the education section. I don't understand Middayexpress' objection about 'actual educational levels'. It doesn't include stats if that is what you mean, but it is still good material. BrumEduResearch (talk) 17:50, 18 February 2015 (UTC)
 * Remarkable how this newly registered account just so happens to agree with Larry in his absence . Middayexpress (talk) 18:14, 18 February 2015 (UTC)
 * I don't know; people have agreed with me before and I wouldn't be surprised if they did so again! Cordless Larry (talk) 00:47, 19 February 2015 (UTC)
 * Remarkable indeed. Middayexpress (talk) 01:23, 19 February 2015 (UTC)

Question
A request was made at the third opinion noticeboard for a third opinion on sources about the level of education of Somalis in the United Kingdom. The threaded discussion is lengthy, and I haven't read through all of the comments, let alone browsed the sources. Can the two editors, User:Middayexpress and User:Cordless Larry, please each summarize concisely (in one paragraph each) what they are asking? I'd like to help, but the comments are too long. Robert McClenon (talk) 23:12, 17 February 2015 (UTC)


 * Sure - thanks for offering to help. I'm suggesting that we make a couple of additions. First, I propose that we include information on the education level of the Somali adult population in the UK based on two sources: average leaving age from this 2007 report and level of qualifications from this 2008 report. The qualification levels need to be caveated with the fact that some foreign qualifications are hard to categorise using the Labour Force Survey categories, as stated in the source.
 * Second, I propose adding the following paragraph: "According to a 2008 report by the Institute for Public Policy Research, most early Somali labour migrants to the UK were literate in English or Arabic, but only a small minority had completed secondary education. When refugee movements started to increase from 1988 onwards, a greater proportion of the refugees had completed secondary education than previous Somalis migrants, and some held university degrees. The report notes that the educational profile of Somali migrants arriving in the UK has changed again since the arrival of this early wave of refugees. Many educational institutions in Somalia have been destroyed by the war and while some have reopened, many young Somali migrants will not have had the chance to attend university and "are likely to have had a very interrupted education or none at all"".
 * I hope that helps. Cordless Larry (talk) 23:16, 17 February 2015 (UTC)

As explained above, the only thing the 2008 paper claims with regard to the actual educational levels of Somalis is the following phrase: "Bangladesh-born and Somalia-born populations were most likely to have no qualifications at all and least likely to possess qualifications at Level Three and above (A-level equivalent)". However, this blurb is contradicted by another 2007 report by the same IPPR, which indicates that such LFS-derived educational estimates are inaccurate, as they underestimate the actual qualifications of immigrants :
 * "The LFS includes questions on qualifications based on equivalent National Vocational Qualification (NVQ) levels, but using these variables to assess the qualification levels of immigrants is problematic since, due to difficulties assessing the UK equivalent of overseas qualifications, foreign qualifications tend to be characterised under a general ‘other qualifications’ heading. As Manacorda et al (2006) note, when native-born respondents report holding ‘other qualifications’, this is generally indicative of very low-level qualifications, whereas when immigrants’ qualifications are classified in this group they are often of a much higher level. This results in the qualification levels of immigrants being under-estimated. In order to avoid this problem, we have used the measure preferred by Dustmann et al (2007): the average age at which each of our groups left full-time education."

Similarly, a 2002 study by Goldsmith's College similarly indicates high levels of educational qualifications for Somali immigrants during the period. The study exclusively aimed to obtain a profile of the professional capacity of Somali nationals living in the UK. In conjunction with 26 Somali community organizations, the researchers interviewed respondents in seven different cities across Britain, and found that:
 * "59% had a qualification from the Somali region, and 9.5% from outside Britain or the Somali regions. 11.5% had a degree or higher degree from the Somali regions compared to 3.5% from beyond Britain and the Somali regions. Since being in Britain, 14% had gained a degree and 12.5% were studying for a degree."

The study also notes that "despite considerable skills and experience brought to Britain through previous employment, 'occupational downgrading' and 'segregation' was prevalent amongst respondents". It concludes that: "the research found that the respondents were highly skilled and educated, often having achieved a large amount of employment experience". Thus, as the 2007 IPPR paper explains, the LFS-derived assessment is indeed underestimating the actual qualification levels of the immigrants. So everything else on the purported educational levels of Somali immigrants in that paragraph by default revolves around this dubious claim. The info should instead therefore be sourced to the Lambeth local authority, which is more factual and official. Middayexpress (talk) 23:32, 17 February 2015 (UTC)
 * I'm going to have to pass on this issue. I don't entirely understand the issue, and neither of the summaries is short enough to get the point across to me.  Maybe you need someone else's third opinion.  Thank you for trying a third opinion.  Robert McClenon (talk) 15:43, 18 February 2015 (UTC)
 * Understood, and thank you for your time. Middayexpress (talk) 16:28, 18 February 2015 (UTC)

Somali pupils
Right, I said I'd come up with some suggested text for the article based on the new source that I found. Here goes:


 * No reliable nationwide statistics are available on the educational attainment of ethnically Somali pupils in the United Kingdom. Central government does not collect educational attainment data for specific migrant groups. Individual schools and local authorities can collect data using extended ethnicity codes, which provide a higher level of detail than the standard classification used in the UK. However, this is not the case in all local authority areas, including some ethnically diverse ones. As Rutter states, "It should be noted that the central government is extremely reticent to carry out a national analysis of educational outcomes by extended ethnicity codes, and spending cuts mean that fewer local authorities are now undertaking such analysis". However, Demie adopts an alternative approach to understanding the educational attainment of minority groups. He notes that while nationwide data by ethnicity is not available, since 2007 local authorities have been required by government to collect data on the main language spoken in the home of pupils, where that language is not English. He presents data from the National Pupil Database on pupils in schools in England who completed their GCSEs in the summer of 2012. Amongst the 2,748 pupils categorised as Black African whose home language was recorded as Somali, 47 per cent achieved five or more GCSEs including in maths and English at grades A* to C. This compares to 58 per cent of all Black African pupils and an overall average for pupils in England of 59 per cent.

There's more information in that source that we could usefully use as background material for the article, but I thought I'd stick to the above for now. We can expand the section further if there is agreement on adding this text. Cordless Larry (talk) 09:49, 6 February 2015 (UTC)
 * Well I have no issue with the above text, I think it is suitable to be addedNograviti (talk) 10:13, 6 February 2015 (UTC)
 * @Larry, I've not managed to access all citations. The second half of the paragraph reads very well, however, I think there should be much less commentary on the trials and tribulations of data supply which currently occupies the first half of the paragraph, particularly as well because it distracts the reader from getting to the meat of the paragragh. It would be enough to say there are problems and limitations reported by commentators on accessing nationally-based data but useful  (or sound) data based on 'pupil database returns' which are standardised across LA's is available. I do think there is more valuable information on the rise of achievement of the Somali community over the last 10 years or so and on barriers to achievement refered to in reference 1 which could be included in the section, particularly as by far the substantial majority of Somali's live in London where the study was undertaken. Tmol42 (talk) 14:59, 6 February 2015 (UTC)
 * Thanks Tmol42. To be clear, I'm not suggesting that this be the totality of the section. I agree that we should make use of the material on London as well. That's already in the article, hence why I was focusing on this part for now, but we need to think about how they fit together as I've used some material already in the article in the passage above. I sort of agree about the data supply discussion, but also feel that it is necessary to explain why the data is about language groups rather than ethnic groups per se. Any suggestions on how that could be explained more briefly would be gratefully received. Cordless Larry (talk) 15:23, 6 February 2015 (UTC)
 * The passage could certainly be condensed to focus on the main figures. Also I think we are all in agreement here that the intention is not portray Somali academic attainment in the UK negatively, but merely provide accurate national figures and show improvements in average attainment where possible.Nograviti (talk) 15:41, 6 February 2015 (UTC)
 * Yes, it would be good to show how things have changed over time. Middayexpress has been opposed to using historical statistics (which are obviously required if we want to demonstrate trends over time) if there are more recent figures available, but perhaps he will reconsider? Cordless Larry (talk) 15:52, 6 February 2015 (UTC)
 * Maybe the sentence starting "As Rutter states..." could be shunted to a footnote? That would help us get to the data more quickly. Cordless Larry (talk) 15:56, 6 February 2015 (UTC)
 * That's what WP:SCHOLARSHIP indicates. On this point, Maunus wrote that "outdated means that it has been superceded by more recent information[...] information about 2001 is not automatically outdated and as long as it is the most recent and it is specified that it is from that date it is not problematic to use it[...] it requires a reliable source to state that the numbers for 2001 are no longer current[...] we cannot simply dismiss it because we assume them no longer to be current." Middayexpress (talk) 16:44, 6 February 2015 (UTC)
 * To clarify, are you saying that you agree we can use old figures to compare with newer ones to demonstrate how Somali pupils' performance has improved, or that you don't think we should do that? Cordless Larry (talk) 16:57, 6 February 2015 (UTC)
 * I am saying the same thing as the post above from 23:28, 5 February 2015 (UTC). Middayexpress (talk) 18:38, 6 February 2015 (UTC)
 * Can you not just answer my question? Do you think we should show how Somali pupils' performance has improved over time or not? Cordless Larry (talk) 18:44, 6 February 2015 (UTC)
 * I already did answer that: Maunus wrote that such material is still relevant if it hasn't been superceded by anything. This is apparently why you insisted on that 89.3% figure from 2001 for the number of Muslim adherents. Middayexpress (talk) 20:31, 6 February 2015 (UTC)
 * That doesn't specifically address whether new and old data can be compared. Anyway, they obviously can be, as is pointed out by another editor below, so it's fine to demonstrate how the performance of Somali pupils has improved over time. Cordless Larry (talk) 22:37, 6 February 2015 (UTC)
 * It's okay if the paper itself does that and its figures are accurate and representative. However, if a Wikipedian tries to string together different sources to create a new narrative of his/her own, it is WP:SYNTHESIS. Middayexpress (talk) 16:11, 7 February 2015 (UTC)
 * Great. That's what I was after. What I had in mind was this article, which mentions that in Lambeth the proportion of Somali pupils getting five good GCSEs rose from 11 per cent to 61 per cent from 2007 to 2012. Cordless Larry (talk) 16:40, 7 February 2015 (UTC)
 * That is inadequate. It should be sourced to the official Lambeth Research and Statistics Unit report, which, unsurprisingly, has a different structure . Middayexpress (talk) 19:01, 7 February 2015 (UTC)
 * Yes, we could use that source. I don't know what policy says that it's preferable to the Economist in this instance, but I'm happy to use it. My only concern would be that people might question why we've singled out Lambeth (could be seen as OR), whereas if we cite the Economist then we can point to their choice of example. Cordless Larry (talk) 19:08, 7 February 2015 (UTC)
 * The Economist journalist again cites original research, which he/she falsely attributes to the Lambeth local authority. The fact remains that the local authority does not use the scheme claimed in that link, but instead a different one as can be seen on its relevant Table 4. Even if this had been accurately rendered, "available, academic and peer-reviewed publications, scholarly monographs, and textbooks are usually the most reliable sources" per WP:SCHOLARSHIP. Middayexpress (talk) 19:17, 7 February 2015 (UTC)
 * The only difference is between 11 and 10 per cent for 2007 (a rounding mistake?), but I'm happy to go with the Lambeth source (although I remain concerned that someone might question why we've chosen Lambeth as opposed to any other LA, but let's see). Cordless Larry (talk) 19:26, 7 February 2015 (UTC)
 * That's not the only difference between them. Middayexpress (talk) 20:55, 7 February 2015 (UTC)
 * OK. Care to explain the others? Cordless Larry (talk) 21:00, 7 February 2015 (UTC)
 * The columns, for one. Middayexpress (talk) 21:06, 7 February 2015 (UTC)
 * The Economist presents the information in text form: "In Lambeth, a borough of London, 61% of Somali pupils got five good GCSEs last year, up from 11% in 2007". The Lambeth report includes it in a table, with the columns representing the different years from 2006 to 2012. So? Cordless Larry (talk) 21:13, 7 February 2015 (UTC)
 * That's not all the journalist writes. He embellishes the first column, much like the other one. Middayexpress (talk) 22:29, 7 February 2015 (UTC)
 * Ah, so it's other aspects of the article you don't like. I was only suggesting using it for that quote. As I said though, I'm happy to use the Lambeth source, especially since the Economist seems to have got the 2007 figure wrong by 1 percentage point (assuming the Lambeth source is correct, which seems reasonable). Cordless Larry (talk) 22:32, 7 February 2015 (UTC)
 * It's not a matter of liking, but of mislabeling and inaccuracy. Middayexpress (talk) 23:16, 7 February 2015 (UTC)

Larry, the paragraph above is unsatisfactory. First, that no reliable nationwide statistics are available on the educational attainment of ethnically Somali pupils in the United Kingdom should be sourced to an official governmental body; in this instance, to the Lambeth Research and Statistics Unit ("No reliable national Somali data is available or collected by DCSF" ). Second, Rutter's discussion of various ethnicity codes is not on why there are no reliable nationwide statistics available on the educational attainment of ethnically Somali pupils, but instead "about integration". As such, it is irrelevant here. Lastly, Demie does not use a racial approach in his paper, but rather a linguistic one. What he writes is that "Somali and Lingala speakers tend to have very low attainment compared to other groups", and provides a 47% GCSE for Somali students. In his more comprehensive 2008 paper exclusively on Somali students, he likewise distinguishes the Somali pupils from the "Black African" students. Demie explains that :
 * "The issue of Somali underachievement is complicated by the problem with categorisation of ‘Somali’ which is broadly defined nationally as African. As a result of the lack of data there were various limitations in past research into Somali underachievement in British schools. The absence of national comparative data which identifies patterns of children of Somali origins, places serious constraints on effecting targeting policy and practice developments at national and local level. However, recently a number of London Local Authorities with high Somali school populations began monitoring and collecting data which has provided an interesting example in research evidence."

The Lambeth Research and Statistics Unit assertion above on Somali pupils should therefore instead be noted in place of the Rutter synthesis. Attempted racial categorization is also inappropriate for this population, as the IPPR explains: "Ethnic categories such as ‘black African’ fail to capture the differences between those born in countries such as Zimbabwe, Nigeria, Kenya and Somalia[...] Comparing ethnic groups with country-of-birth groups is also problematic because some countries’ populations may contain multiple ethnic groups". Middayexpress (talk) 16:44, 6 February 2015 (UTC)
 * Sorry, I find it quite difficult to follow your argument there. You say that the point about no nationwide figures being available should be sourced to the Lambeth Research and Statistics Unit. If you read the references, you'll see that that's exactly what I've done (although it isn't a requirement of WP:RS that the source be government). The point about the "Black African" category failing to differentiate between Somalis, Nigerians, etc. is exactly why he focuses it down to Somali speakers. How many Nigerians speak Somali at home? Rutter's statement is not irrelevant; it's an explanation of why national statistics by ethnicity aren't available. There is no "attempted racial categorization"; Demie uses a linguistic categorization, as you yourself point out. Cordless Larry (talk) 16:52, 6 February 2015 (UTC)
 * Larry, In respone to your request @15:56 above my condensed version below. Agree a note of the context could be relegated to a footnote.
 * Although no reliable nationwide statistics are available on the educational attainment of ethnically Somali pupils in the United Kingdom [1][2] some schools and local authorities collect data using extended ethnicity codes. Additionally, since 2007 local authorities have been required by government to collect data on the main language spoken in the home of pupils, where that language is not English as part of the National Pupil Database.  In the summer of 2012, of the 2,748 pupils categorised as Black African whose home language was recorded as Somali, 47 per cent achieved five or more GCSEs including in maths and English at grades A* to C. This compares to 58 per cent of all Black African pupils and an overall average for pupils in England of 59 per cent.[5]
 * Btw I have failed to understand what is meant in all the chat coming back to you since then. is it just me or is me that finds it all totally opaque or worse?Tmol42 (talk) 17:48, 6 February 2015 (UTC)
 * Thanks. I'm happy with that version. I think Midday needs to try to express his objections to using these figures more succinctly, because I don't understand the reply above either. Cordless Larry (talk) 17:52, 6 February 2015 (UTC)
 * I already did in the link above (where you didn't seem to have any trouble understanding me). Middayexpress (talk) 19:10, 7 February 2015 (UTC)
 * Larry, that no reliable nationwide statistics are available on the educational attainment of ethnically Somali pupils in the United Kingdom should be sourced to an official governmental body, as it's the government that gathers this data. The Lambeth Research Statistics Unit explains that no such data is collected by the DCSF, so it should be sourced to it. Further, Rutter writes that local authorities in England analyse data using broad ethnic categories, but they can also use extended ethnicity codes. In reality, this data processing varies greatly between local authorities (e.g. Camden ). Her explanation as to why there is a dearth of nationwide stats on Somali students is also inconsistent with the governmental Lambeth Research and Statistics Unit's official explanation (shown above). The official governmental explanation is therefore more reliable here. Additionally, there is indeed attempted racial categorization in that draft paragraph above. Demie however does not structure his paper in this way; he instead uses a linguistic scheme. That aside of his is also from his 2008 report exclusively on Somali students, where he distinguishes his Somali students from his "Black African" pupils. It's not from his 2014 paper as you appear to assume . At any rate, this discussion could use some actual Somali participants; I'll see to that shortly. I'll also write Demie later on today and link him to the page. His insight would certainly be valuable, given his position as probably the authority on Somali student attainment. Middayexpress (talk) 18:38, 6 February 2015 (UTC)
 * It would be great to get Demie's input, but I wonder how Wikipedia would treat his response. If it's by e-mail, I presume that can't be used as a source as it won't be published. Cordless Larry (talk) 22:45, 6 February 2015 (UTC)
 * I had no intention of actually citing his words; I'm sure Demie would be more than happy though to point me to publications for that. What his input will do is definitively clarify the actual state of Somali student attainment. I'm also curious as to what he makes of this discussion. Middayexpress (talk) 18:49, 7 February 2015 (UTC)
 * Yep, me too! Cordless Larry (talk) 19:01, 7 February 2015 (UTC)
 * Argh! It is sourced to Lambeth. See reference 1! Cordless Larry (talk) 18:41, 6 February 2015 (UTC)

Why is the education section only on London. Somalis live in other place to. — Preceding unsigned comment added by 85.255.234.249 (talk) 23:25, 6 February 2015 (UTC)
 * I dont really follow Middayexpress' recent objections, I think we have some consensus on sources for national data on the academic performance of Somalis. I also don't see the merit in the requirement that other Somalis should comment on this issue? Nograviti (talk) 14:23, 7 February 2015 (UTC)
 * There is certainly not consensus for that paragraph, nor unfortunately am I the only objector . Middayexpress (talk) 16:11, 7 February 2015 (UTC)
 * Could you perhaps indicate (briefly) what parts of the paragraph you object to? I can't really understand your longer post above either (and as I pointed out twice, the first bit about using the Lambeth source doesn't make sense as I already do use it). Do you approve of the use of the 2014 journal article, for instance? Please feel free to suggest an alternative wording, too. Cordless Larry (talk) 16:42, 7 February 2015 (UTC)
 * I already did in the link above (where you didn't seem to have any trouble understanding me). Middayexpress (talk) 18:49, 7 February 2015 (UTC)
 * If you mean the user page link, there's lots of discussion there but it's not really clear to me whether you're arguing against my wording, or against use of the 2014 Demie source altogether. Some more clarity from you on this would be great. Cordless Larry (talk) 19:01, 7 February 2015 (UTC)
 * Yes, I mean the page I just linked to. I'll paste the text here if there's any further trouble understanding what is written there. Middayexpress (talk) 19:10, 7 February 2015 (UTC)
 * Most Somalis live in London, so that's where the most representative data has been gathered. Middayexpress (talk) 16:11, 7 February 2015 (UTC)

Alternative wording suggestion
A discussion (see above) has been taking place about how to incorporate data from this journal article into the education section of this article. The source presents data on the GSCE performance of pupils who speak Somali at home. The reason it uses this data is because there is no nationwide data on GCSE results by ethnicity at the level of detail that allows individual ethnic groups to be identified (so Somalis are often subsumed into a bigger "African" category).

This is Tmol42's suggested wording from above:


 * Although no reliable nationwide statistics are available on the educational attainment of ethnically Somali pupils in the United Kingdom [1][2] some schools and local authorities collect data using extended ethnicity codes. Additionally, since 2007 local authorities have been required by government to collect data on the main language spoken in the home of pupils, where that language is not English as part of the National Pupil Database.  In the summer of 2012, of the 2,748 pupils categorised as Black African whose home language was recorded as Somali, 47 per cent achieved five or more GCSEs including in maths and English at grades A* to C. This compares to 58 per cent of all Black African pupils and an overall average for pupils in England of 59 per cent.[5]

Middayexpress has suggested elsewhere this alternative wording (we would follow this with the statistics from Demie, I presume):


 * According to the Lambeth Research and Statistics Unit, no reliable nationwide statistics are available on the size and educational attainment of Somali pupils in the United Kingdom. Data on the students has often been aggregated under a broad continental 'African' variable, which obscures the students' unique charateristics and requirements. This in turn inhibits targeted policy making and practice developments at the national and local level. To redress this, various London Local Authorities, where most Somali pupils attend school, have started gathering and monitoring data on the Somali student community.

Could we get editors' preferences, if they have any, on which of these to use, or comments for improving either of them? Cordless Larry (talk) 08:05, 8 February 2015 (UTC)
 * Sure but I thought you offered to drop the "Black African" stuff? Anyways, I recommend Middayexpress'. AcidSnow (talk) 14:23, 8 February 2015 (UTC)
 * We could drop the second mention of "Black African" in Tmol42's wording, but to drop the first one would be to misrepresent Debie's data. I'm happy to drop the second mention ("compares to 58 per cent of all Black African pupils"), but I didn't include that here because I didn't want to mess with Tmol's suggestion without his/her awareness. It's more the first part of the paragraph that I think we need input on. Cordless Larry (talk) 14:29, 8 February 2015 (UTC)
 * If I may butt in here re my redraft, I do think the 'Black African' is relevant as it arrises as it is a subset of the Black/African/Carribean/British ethic group from the UK census. There are I recall also not insignificant numbers of Somalis who self-classified themselves as White, or Mixed, or Asain in the results of the 2011 census. I am also aware that the term is sometimes queried as to whether it should apply to 'north Africans'.Tmol42 (talk) 14:51, 8 February 2015 (UTC)
 * Just to explain my comment about the danger of misrepresenting Demie's data for those who don't have access to the article, the title of the table that the data is from is: "Table 2. GCSE performance of Black African pupils by language spoken at home". Cordless Larry (talk) 14:58, 8 February 2015 (UTC)

He did, AcidSnow. Larry asserted that he wasn't interested in attempting to racially classifying this population after we pointed out Somalis' general self-perception, policies discouraging this (viz. WP:CATEGRS and Naming conventions (ethnicities and tribes)), and actual ancestry. He also wrote that my suggested alternative wording above based on the Lambeth Research and Statistics Unit "looks good". Middayexpress (talk) 15:02, 8 February 2015 (UTC)
 * Yes, I don't think there's anything wrong with your wording, although it is a bit long. It's not for me to decide though - we need to reach wider consensus. I think I still prefer Tmol42's wording as it is a bit shorter and leads into the Demie statistics more naturally. Cordless Larry (talk) 15:08, 8 February 2015 (UTC)

The racial phrasing in it is gratuitous and synthesis. Demie certainly does not indicate this in his Lambeth Research and Statistics Unit report, where he actually explains why there's a dearth of nationwide stats on the Somali student population's attainment and size. Even if he had, it would still be a breach of said policies given this population's general self-perception and actual ancestry. Here's what Demie actually writes on this, which my alternative wording above faithfully adheres to: "The issue of Somali underachievement is complicated by the problem with categorisation of ‘Somali’ which is broadly defined nationally as African. As a result of the lack of data there were various limitations in past research into Somali underachievement in British schools. The absence of national comparative data which identifies patterns of children of Somali origins, places serious constraints on effecting targeting policy and practice developments at national and local level. However, recently a number of London Local Authorities with high Somali school populations began monitoring and collecting data which has provided an interesting example in research evidence". Middayexpress (talk) 15:02, 8 February 2015 (UTC)
 * So what is the problem? Can we simply quote him even though it's too long? It seems to make it neutral for both parties. AcidSnow (talk) 15:21, 8 February 2015 (UTC)
 * Yeah, that would most neutrally resolve the matter. The passage is a bit long for that, though; better to paraphrase it, which I've just done. Middayexpress (talk) 15:35, 8 February 2015 (UTC)
 * Yep, all my contributions are free of charge, thanks for acknowledging that!Tmol42 (talk) 16:55, 8 February 2015 (UTC)
 * Gratuitous as in unnecessary. Middayexpress (talk) 17:22, 8 February 2015 (UTC)
 * If anyone wants to see the table in the Demie source (it might help explain the "Black African" wording), here is a screenshot. Cordless Larry (talk) 22:31, 8 February 2015 (UTC)



Larry, I asked you to link externally to a screenshot, not to upload a copyrighted file onto Wikipedia. This is quite irrelevant anyway since Demie's actual explanation for why no reliable nationwide statistics are available on the size and educational attainment of Somali pupils in the United Kingdom is from his report on Somali students (where he distinguishes his Somali students from the Black African pupils ), not the 2014 paper on language. It doesn't really matter though since, as it turns out, Demie does indeed provide a nationwide GCSE figure for 2007 for Somali students in England (not Somali language speaking pupils, which would include Bantus and other ethnically non-Somali individuals). It's in his Lambeth Research and Statistics Unit paper for Portuguese students; the figure for Somali pupils is 42% ("Table 2 shows KS2 and GCSE results for each main ethnic group at national level" ). Thus, this is the actual paper to cite for the latest available national GCSE for this population. Middayexpress (talk) 16:10, 9 February 2015 (UTC)
 * The screenshot discussion above was about an IPPR report, not this paper. Use of this doesn't break copyright - I asked for advice at Media_copyright_questions. Cordless Larry (talk) 16:37, 9 February 2015 (UTC)
 * Also, there's no table 2 in the article you've linked to there, and that quote doesn't appear in it either. Wrong link, perhaps? Cordless Larry (talk) 16:41, 9 February 2015 (UTC)
 * Actually, this screenshot was from Demie's actual paper on Somali students . And yes, he certainly does include a national GCSE for Somali pupils in his Table 2; the figure is 42% ("Table 2 shows KS2 and GCSE results for each main ethnic group at national level" ). Thus, this is indeed the actual paper to cite for the latest available national GCSE for this population. Middayexpress (talk) 17:15, 9 February 2015 (UTC)
 * OK, you linked to the wrong paper at the end of the quote in your comment above then. Yes, that appears on first sight to be a national figure. That's confusing when you consider Demie's claims in other papers that no such national figure exists for Somalis. The answer appears to lie in Table 3 here. Note that the data is exactly the same as in the Portuguese pupils paper. There is a footnote next to the Somali figure, though, which states: "The Somali figure here is the average of over 28 London Local Authorities with over 2500 Somali pupils in primary school, and 1200 pupils in secondary schools. No Somali national data is available". So what they're doing is comparing the Somali figure for London to the national figure for other, broader groups. However, given that they've made that comparison and that most Somalis are based in London, perhaps we could indeed use the 42 per cent figure in the article? Cordless Larry (talk) 17:24, 9 February 2015 (UTC)
 * Yes, that's quite possible. I think we should note both Demie's 34% national GCSE for 2006 and his 42% national GCSE for 2007 to show how much the latest available national GCSE figure for Somali students changed within a year and the direction it was moving in. Middayexpress (talk) 18:26, 9 February 2015 (UTC)
 * The 34 per cent from Figure 1? The only problem there is that they're not directly comparable as that figure is based on 10 local authorities in London, not the 28 used to calculate the 2007 figure. We could caveat the statistics with this, though. Cordless Larry (talk) 18:31, 9 February 2015 (UTC)
 * Section 4.2.2 of this paper surveys a variety of figures, include Demie's. Another alternative is to quote/cite that. Cordless Larry (talk) 18:35, 9 February 2015 (UTC)
 * It's unclear where that paper is getting its 2005 national GCSE figure from, as the 2007 DFES report itself doesn't actually give any such figure for Somali students . It also ignores the many local authorities that don't process their data in that way, and uses a tiny sample of 98 Somali students. Given this, I think we should link directly to Demie's actual papers, noting both his 34% national GCSE for 2006 [ and his 42% national GCSE for 2007, with a caveat as to the number of London local authorities that were used to calculate these national GCSE figures. [[User:Middayexpress|Middayexpress]] (talk) 19:31, 9 February 2015 (UTC)
 * OK, agreed. Cordless Larry (talk) 19:54, 9 February 2015 (UTC)
 * Alright, done. Middayexpress (talk) 22:28, 9 February 2015 (UTC)
 * Good. I just made some small changes of wording ("GSCE score" isn't quite right - it's the proportion gaining five or more good grades at GCSE) and also removed one reference where the statistic appeared in another, already cited one that is specifically about Somalis. Cordless Larry (talk) 22:52, 9 February 2015 (UTC)
 * Sorry to throw yet another spanner in the works, Middayexpress, but it seems the 2006 figure is for the proportion gaining five A*-C grades regardless of subject, whereas the 2007 figure is the proportion gaining five A*-C grades including maths and English, so they're not quite measuring the same thing. The proportion of the latter is always lower than the former, so this probably underestimates the improvement (although the difference in the number of LAs complicates the comparison). Cordless Larry (talk) 23:23, 9 February 2015 (UTC)
 * They are measuring the same thing; please see Table 6 & Figure1 . Middayexpress (talk) 23:39, 9 February 2015 (UTC)
 * You're right. I looked into the sources of the broad ethnicity group data and got the idea that one included maths and English, but checking again I must have been mistaken. The proportions in that table and chart are all too high to be the maths and English figures. Phew! Cordless Larry (talk) 00:08, 10 February 2015 (UTC)
 * No prob. Middayexpress (talk) 00:53, 10 February 2015 (UTC)
 * A couple of minor suggestions for rewording here, to improve the clarity of the text. It currently reads: "Based on the GCSE averages from 10 London local authorities, the Lambeth Research and Statistics Unit estimated that 34 per cent of Somali students in England gained five or more GCSEs at grades A*–C in 2006. This proportion had increased to 42 per cent by 2007, based on a larger sample of the averages of 28 London local authorities". First, the "in England" is a bit confusing, since that data is clearly for the 10 London LAs mentioned earlier in the sentence, not for England as a whole (as the article explains, such national data is not available). Second, the "had increased" bit is misleading since it doesn't compare like with like. It's a comparison of data from 10 LAs with data from 28, and the source itself doesn't claim that this is proof of an increase. Given the wide variation in pupil performance across London, it might just be that the additional 18 LAs already had better-performing Somali pupils. I suggest something along the lines of "This compares with a figure of 42 per cent in 2007...". Cordless Larry (talk) 22:06, 17 February 2015 (UTC)
 * The "in England" bit is necessary since this is what Table 6 actually indicates, and without it the estimate does not appear to be national, when it is in fact intended as such. The "had increased" bit is thus not misleading since both tables are captioned GCSE Attainment by Main Ethnic Groups in England 2007; it also indicates the actual direction of the change in GCSE average. Middayexpress (talk) 22:21, 17 February 2015 (UTC)
 * The footnotes in the source make it clear that the Somali figures are not for England as a whole, but for some London LAs. In what sense does that make them "for England"? Cordless Larry (talk) 22:27, 17 February 2015 (UTC)
 * Also, the text of the report states "Table 3 shows KS1, KS2, KS3 and KS4 results for each main ethnic group at national level including aggregated data for selected LAs in London which are noted as having over two thousand Somali pupils in their schools". That makes it clear that the Somali data is for London LAs, and the broad ethnic groups data for England as a whole. Our wording should reflect this. Cordless Larry (talk) 22:48, 17 February 2015 (UTC)
 * As for "both tables are captioned GCSE Attainment by Main Ethnic Groups in England 2007", one of the figures is from a chart for 2006, not a table for 2007. Cordless Larry (talk) 22:29, 17 February 2015 (UTC)
 * See Figure 1. Middayexpress (talk) 22:56, 17 February 2015 (UTC)
 * Yep, that's what I'm referring to. The Somali data is for London alone, as made clear by the footnote and the text that introduces the chart. Cordless Larry (talk) 23:02, 17 February 2015 (UTC)
 * You're looking at the wrong Figure 1 (there are several therein). It's the one titled "Somali and Main Ethnic Groups GCSE Performance in England, 2006", on page 6. Middayexpress (talk)
 * That's what I'm looking at. The footnote states: "The Somali figure here is the average of 10 Local Authorities in London (Demie, 2006). No reliable national Somali data is available or collected by DCSF". The data is for 10 London LAs, not for England as a whole. The other items in the figure are for England as a whole. Cordless Larry (talk) 23:12, 17 February 2015 (UTC)
 * Indeed. Its extrapolated for England as a whole because there is no nationwide data available; see Table 6 titled "GCSE National Performance in England by Ethnicity and Gender – 2007" ("No national Somali data available. Data here refers to London Local Authorities." ). Middayexpress (talk) 23:47, 17 February 2015 (UTC)
 * Yes, so why are we suggesting in our reporting of the data that it applies to England as a whole? Also, do you agree that the source makes no claim about an improvement between 2006 and 2007? It just presents different data on the two years. Cordless Larry (talk) 23:48, 17 February 2015 (UTC)
 * The Lambeth local authority itself already does this when it compares the Somali students GCSEs with the national (not London) GCSEs of other pupils. That's what the "in England" in each table's title means. Middayexpress (talk) 23:59, 17 February 2015 (UTC)
 * Yes, it compares London data for Somalis with England data for other groups. We should reflect this in our wording. That's what I'm suggesting we do. Cordless Larry (talk) 00:01, 18 February 2015 (UTC)
 * Actually it compares Somali data culled from London but extrapolated for England with nationwide data for other populations. That is why the table as a whole is titled "in England". The wording captures this. Middayexpress (talk) 00:22, 18 February 2015 (UTC)
 * It's compared, not extrapolated. The source doesn't claim the figures are representative of England as a whole. Our wording is confusing as it says it's for London, and then for England. If you insist on that, I suppose it will have to remain confusing. How about the "improvement" wording? That's not supported by the source. Cordless Larry (talk) 00:24, 18 February 2015 (UTC)
 * Yes it is indeed extrapolated nationally (see page 17 ). Also, the actual wording notes that "this proportion had increased to 42 per cent", which is also extrapolated nationally (see Table 2 ). Middayexpress (talk) 01:03, 18 February 2015 (UTC)
 * Yes, as previously discussed, that table in the report on Portuguese pupils is the same as the one in the Somali pupils report, without the footnote. Neither report states that GCSE performance has increased, because the 2006 and 2007 figures are based on very different sized samples. Cordless Larry (talk) 06:43, 18 February 2015 (UTC)
 * Also Larry, kindly refrain from following me around on unrelated pages (see WP:HOUNDING). Middayexpress (talk) 01:19, 18 February 2015 (UTC)
 * Sorry, where have I hounded you? I've been discussing a similar issue with you at Talk:Nigerian British, but I have a longstanding interest in that page. Cordless Larry (talk) 06:41, 18 February 2015 (UTC)
 * You followed me yesterday onto an ip's talk page and the Mogadishu page, discussions which have nothing to do with you (other than my presence there of course) . Kindly stop it. Middayexpress (talk) 16:28, 18 February 2015 (UTC)
 * Discussions about the Mogadishu article have nothing to do with me? I'm as free to contribute as anyone. Also, I wasn't aware that I could hound you by agreeing with you, which is what I did there! Cordless Larry (talk) 16:31, 18 February 2015 (UTC)
 * Hounding: "Wikihounding is the singling out of one or more editors, and joining discussions on multiple pages or topics they may edit or multiple debates where they contribute, in order to repeatedly confront or inhibit their work. This is with an apparent aim of creating irritation, annoyance or distress to the other editor. Wikihounding usually involves following the target from place to place on Wikipedia." Again, please stop it. Middayexpress (talk) 17:35, 18 February 2015 (UTC)
 * Agreeing with your argument does not constitute inhibiting your work or (I presume) annoying or distressing you. If you have been distressed by my agreeing with you and providing a source that supports your argument, please take it up on my talk page rather than here, or report me and have an admin look into it. Cordless Larry (talk) 23:59, 18 February 2015 (UTC)
 * It makes no difference whether or not you agreed with my argument. Don't follow me around again. Middayexpress (talk) 00:40, 19 February 2015 (UTC)
 * Sorry, I won't be told what articles I can and can't edit by you. As I said, please report me if you feel that I have broken the rules. Cordless Larry (talk) 00:45, 19 February 2015 (UTC)
 * You both kind of missed the point each of you made. Midday isn't forbiding you from editing articles Larry, he just doesn't want you to follow him around. Also that in this instant you did follow him. As for you Midday, you should at least be thankful he gave you support when you never asked him to. P.S the text gets smaller and smaller and smaller as you read lol. AcidSnow (talk) 01:02, 19 February 2015 (UTC)
 * Yes, I did follow Midday there, but my actions don't meet the definition of hounding. I repeat, though, that this is not the place to discuss this. Please report me to an admin if you think I have acted inappropriately, Midday, rather than continuing to accuse me here. Cordless Larry (talk) 08:07, 19 February 2015 (UTC)

I scanned the above so sorry if I missed anything. Seems to me you just need to describe the data. Leave out things like 'improve' unless the source says this exact word. Readers can be left to make their own minds up about that easily enough. BrumEduResearch (talk) 15:01, 18 February 2015 (UTC)
 * Thanks. Although "improve" wasn't part of the wording, I agree that it's best to just neutrally relay the data. Middayexpress (talk) 16:28, 18 February 2015 (UTC)
 * Agreed, but is it possible that BrumEduResearch meant "increase" when they said "improve"? Cordless Larry (talk) 16:39, 18 February 2015 (UTC)
 * Sorry, that is what I meant to write. The source does not even include the 2006 and 2007 figures in the same table. They're not comparable because they are based on very different samples. I would just report the 2007 figure for Somalis in the 28 London authorities and compare with the England figure for all pupils from the same table. That cuts the massive section down somewhat too. BrumEduResearch (talk) 17:43, 18 February 2015 (UTC)
 * The 2006 and 2007 figures are from the same Lambeth local authority, so they are indeed comparable. I also couldn't help but notice that you just registerd this account. Remarkable timing there . Middayexpress (talk) 18:08, 18 February 2015 (UTC)
 * I replied your timing comment below. I'm not sure what I have to do as a new/returning editor. The figures come from the same source but it doesn not compare them as they cover different areas. BrumEduResearch (talk) 18:12, 18 February 2015 (UTC)
 * The tables are both extrapolating the GCSE figures for Somali students in England in 2006 and 2007, respectively, based on the averages of various London local authorities. The only difference is that the 2006 average is based on 10 LLAs while the other one is based on 28 LLAs. Also, new and "returning" editors likewise aren't the same thing. Middayexpress (talk) 18:35, 18 February 2015 (UTC)
 * They just compare London Somali with England all pupils. It is easier to just put the stats like that. Your extropolation wording is confusing and not in the source. BrumEduResearch (talk) 18:37, 18 February 2015 (UTC)
 * Unfortunately, the title of the tables themselves indicate otherwise (e.g. ). Middayexpress (talk) 18:52, 18 February 2015 (UTC)

Are you actually dumb or just acting it? Read the footnote. BrumEduResearch (talk) 11:11, 20 February 2015 (UTC)
 * That personal attack is duly noted. And after Gigs specifically asked to remain civil at that. Middayexpress (talk) 17:13, 25 April 2015 (UTC)


 * 3O/Comment I've removed your third opinion request because you have several involved editors, and have already started an RfC.  I will add some thoughts regardless.   Please remain civil and refrain from fighting over petty issues. Keep in mind that this entire lengthy discussion is centered on a small detail of the article.  We do not need to incorporate information from every possible source, especially if sources contradict each other.   It may be sufficient to merely comment that the education level of Somali immigrants is difficult to judge, which can be backed up with sources that say that.  It seems that there is definite consensus that there is a lack of reliable data on this issue, so consider just going with that, rather than spending a lot of time debating on how to present imperfect and potentially flawed data. Gigs (talk) 17:09, 18 February 2015 (UTC)
 * Thanks Gigs; quite sensible advice. Middayexpress (talk) 17:35, 18 February 2015 (UTC)
 * Thanks from me too, Gigs. Actually, I hadn't intended the request to be about this section, but the section above titled Level of education of Somali population in the UK. There have indeed been several contributors to this discussion about Somali pupils' results, but at the time of the request only two editors had been involved in the discussion in the other section, which is about the level of qualifications achieved by Somalis in the UK as a whole (i.e. including adults). Still, perhaps these two topics aren't discrete enough for you to treat them as separate. Cordless Larry (talk) 00:03, 19 February 2015 (UTC)
 * Gigs' remark refers to the education level of Somali immigrants in general. And he is certainly right; there is definite consensus that there is a lack of reliable data on this issue, so it is indeed sufficient to note that it is difficult to judge. Middayexpress (talk) 00:47, 19 February 2015 (UTC)
 * There are plenty of reliable sources on Somali educational underachievement in Britain. BrumEduResearch (talk) 12:10, 19 February 2015 (UTC)
 * If one deliberately searches for underachievement in any given population (as you did), one can find it, including underachievement of British pupils (e.g. ). Likewise, one can also find material on high achievement of pupils, including Somali students (e.g. ). The actual official GCSE results for Somali students, however, are increasingly towards the latter. Middayexpress (talk) 16:59, 19 February 2015 (UTC)
 * In individual LAs, yes. But at the broader scale, there is consensus that Somali pupils underachieve, for a variety of very understandable reasons. The Lambeth report which you like so much, Midday, states "Yet the achievement of Somali heritage pupils lags far behind the average achievement of the majority of their peers and the gap is growing at the end of primary and secondary education". Cordless Larry (talk) 07:41, 20 February 2015 (UTC)
 * That Demie report is reliable. He's an authority on the topic. BrumEduResearch (talk) 11:08, 20 February 2015 (UTC)
 * Larry, as you are well aware, that Lambeth local authority also indicates that "research evidence also show Somali pupils are highly motivated and once Somali children reach a competent level of English they do better than English, Scottish and Welsh", and that "the evidence is that once Somali children reach a competent level of English, they forge ahead in their learning and can reach the highest standards". The actual latest official GCSEs of Somali students show this too, with the Somali pupils performing on par with the national average by 2011-2012 per both the Camden and Tower Hamlets local authorities. Middayexpress (talk) 17:07, 20 February 2015 (UTC)
 * Yes indeed, so why do we have to be selective and only present what you see as the positive aspects of the Lambeth report. We could easily include both my quote above and the ones you cite in the article. Cordless Larry (talk) 17:18, 20 February 2015 (UTC)
 * Selective? Hardly (that would be typing in "underachievement" instead of simply "achievement"). The page already notes that "Key Stage 2 (KS2), Key Stage 3 (KS3) and GCSE trend data from several LAs indicated that Somali students on average underachieved compared to the national average of their peers from other ethnic groups." Only it does so in context and explains the reasons for that underperformance (mainly lack of English profiency), as well as what actually happens when these factors are adjusted for (high scores). Middayexpress (talk) 18:34, 20 February 2015 (UTC)
 * The GCSE discussion should be shorter. It is sufficient to say that Somalis underperform and report Demie's statistics for 2007. Then explain that results vary across London. Say they are good in Camden if you want. Then explain the reasons for underperformance. Simples. BrumEduResearch (talk) 11:35, 22 February 2015 (UTC)
 * No... it is not sufficient to say that Somalis underperform without contextualization because that is certainly not all Demie says (as shown above). That is one-sided pov and soapboxing, which is against policy. Middayexpress (talk) 15:47, 22 February 2015 (UTC)
 * That's what I'm suggesting we do: explain the underperformance. We could even do that at the very start of the section if you like. BrumEduResearch (talk) 12:08, 23 February 2015 (UTC)
 * The students' overall performance, as well as the various main factors determining that performance (viz. English proficiency, pupil mobility, teacher & parental engagement) are already noted per the Lambeth local authority. So is what actually happens when those key factors are adjusted for (high scores). Middayexpress (talk) 15:19, 23 February 2015 (UTC)
 * I know. I am suggesting how to make the section shorter because three paragraphs on GCSEs alone is undue. BrumEduResearch (talk) 15:27, 23 February 2015 (UTC)
 * Actually, only the last paragraph is on the actual GCSE scores. The first is on why no reliable nationwide statistics are available on the size and educational attainment of Somali pupils in the UK, and the second is on the pupils' overall performance, the various main factors determining that performance, and what happens when those key factors are adjusted for. All of the data is also official. Middayexpress (talk) 15:56, 23 February 2015 (UTC)
 * Yes but it's all about school education up to GCSE. Nothing about further and higher ed or about adult Somalis' quals. BrumEduResearch (talk) 15:35, 27 February 2015 (UTC)
 * Now you're claiming that it's schooling only up to GCSE, whereas above you claimed (wrongly) that it was on GCSEs alone. At any rate, the professional qualifications past the GCSE level are uncertain, as Gigs made clear. Middayexpress (talk) 16:04, 27 February 2015 (UTC)

Middayexpress, I think you need consensus for your text. It looks like there's more support for @User:Tmol42's wording. BrumEduResearch (talk) 12:43, 25 April 2015 (UTC)
 * Actually, no. Myself and AcidSnow certainly don't support it. It's not particularly relevant anyway, since this page is on residents with Somali ancestry, not on Somali language speakers (which include many non-Somali individuals). Gigs likewise noted there is definite consensus that there is a lack of reliable data on this issue, so it is indeed sufficient to note that it is difficult to judge. Middayexpress (talk) 17:13, 25 April 2015 (UTC)
 * It's cerainly no suppored by Midday or I. AcidSnow (talk) 17:41, 25 April 2015 (UTC)
 * Gigs as well and surely, 26oo. Middayexpress (talk) 18:57, 25 April 2015 (UTC)
 * Wow, I'm tired of this debate. On a purely numerical basis, I think there is greater support for Tmol's suggested wording than there is for the current wording (I count three supporters versus two), but I wouldn't say there's consensus for either. I do think the current section goes into too much detail about what is a relatively minor aspect of the overall topic. Cordless Larry (talk) 18:06, 25 April 2015 (UTC)
 * Thinking about it, this whole section is probably a good candidate for dispute resolution as I don't see the discussion here going anywhere. Cordless Larry (talk) 18:12, 25 April 2015 (UTC)
 * Given the Rfc and Third Opinion, that would be WP:FORUMSHOPPING. You yourself also admitted that the language based GSCEs aren't as reliable, as they aren't predicated on ancestry but rather on language. Middayexpress (talk) 18:57, 25 April 2015 (UTC)
 * When those haven't resulted in consensus, taking it to dispute resolution isn't forum shopping, it's just the next step in the process. 26oo hasn't even commented on this discussion, and as far as I can see, Gigs didn't offer support for all of this highly detailed material about GSCE results in London. Cordless Larry (talk) 19:06, 25 April 2015 (UTC)
 * Consensus is only "ascertained by the quality of the arguments given on the various sides of an issue, as viewed through the lens of Wikipedia policy." Forum shopping is also: "Raising essentially the same issue on multiple noticeboards, or to multiple administrators, is unhelpful to finding and achieving consensus. It does not help develop consensus to try different forums in the hope of finding one where you get the answer you want." That said, Gigs already pointed out that he "removed your third opinion request because you have several involved editors, and have already started an RfC." He also explained that there is definite consensus that there is a lack of reliable data on this issue, so it is indeed sufficient to note that it is difficult to judge. That certainly doesn't support the language based GCSE; quite the contrary. Per WP:NOCON, here's the actual policy on the matter: "in discussions of proposals to add, modify or remove material in articles, a lack of consensus commonly results in retaining the version of the article as it was prior to the proposal or bold edit". That means the incumbent non-language based GCSE material is retained. Middayexpress (talk) 19:59, 25 April 2015 (UTC)
 * To be clear, I wasn't suggesting that Gigs supported Tmol's version, just that Gigs hadn't signaled support for the long and detailed discussion of GCSE results in London. I also still can't see where 26oo supported your version, as you have claimed. Cordless Larry (talk) 20:11, 25 April 2015 (UTC)
 * I didn't write that 26oo supported the non-language based GCSE material. I suggested that he likely would, given his knowledge on Somali related affairs. Gigs likewise wrote that "it seems that there is definite consensus that there is a lack of reliable data on this issue, so consider just going with that, rather than spending a lot of time debating on how to present imperfect and potentially flawed data". That indeed certainly doesn't support the language based GCSE material. This is irrelevant anyway, as the query ended and WP:NOCON is quite clear that the incumbent material in such matters is retained. Middayexpress (talk) 20:59, 25 April 2015 (UTC)
 * We all agree on the fact that there is a lack of reliable data. That doesn't mean there's consensus for all of the material you added about GCSE results in London. Claiming that 26oo "likely would" support you, without them having even commented on this section, is very questionable if you ask me. Cordless Larry (talk) 21:08, 25 April 2015 (UTC)
 * I didn't write that 26oo would likely support me personally either, but rather the non-language based GCSE material. This is because not all Somali language speakers are of Somali ancestry. At any rate, per policy, the default for a lack of consensus is that the incumbent material is retained. Middayexpress (talk) 22:11, 25 April 2015 (UTC)
 * The problem is that it's not really clear what the stable wording is, and several editors oppose elements of the existing focus on London GCSE data. That's why I suggested dispute resolution, because the discussion here seems to be heading nowhere and meanwhile the section is left with a UNDUE tag. Cordless Larry (talk) 23:46, 25 April 2015 (UTC)
 * I realize that there was contention over the non-language based GCSE material. However, that's ostensibly what the Rfc was for. It already expired in March, and per WP:NOCON, the default for a lack of consensus is that the incumbent material is retained. Middayexpress (talk) 00:16, 26 April 2015 (UTC)
 * As suggested at ANI, I will consider requesting mediation on this issue once the other RfC has expired. Cordless Larry (talk) 05:46, 26 April 2015 (UTC)
 * Ok; but as also noted there by Robert, it's a voluntary process. Middayexpress (talk) 15:31, 26 April 2015 (UTC)