Talk:Ethnic heterogeneity
From Metapedia
Ethnic heterogeneity associated with lower life satisfaction - methodological question
Hello Upplysning, again welcome here at Metapedia. I am reading this new article "Ethnic heterogeneity" with interest and I appreciate that you refer to scientific publications as support for the statements in the article. In the section entitled "Life satisfaction", the following statement is made:
- "A 2004 study in the UK found that higher ethnic heterogeneity in an area was associated with lower life satisfaction .[12]"
The reference [12] is to "Bobby Duffy. Life satisfaction and trust in other people. March 2004. Mori social research institute.", but the link http://ipsos-rsl.com/DownloadPublication/1226_sri_life_satisfaction_and_trust_in_other_people_122004.pdf seems to be dead.
I wanted to check this publication, because I want to know if the assocation was statistically significant (the p-value) and for which variables adjustments were made. I will try if I can somewhere find a pdf of this publication. In the meantime, could you please add this information to the article? Many thanks in advance, this will help with interpreting/understanding the association. (Galileo 22:39, 14 November 2014 (CET)).
- This should be a working link: https://www.ipsos-mori.com/researchpublications/publications/1226/Life-satisfaction-and-trust-in-other-people.aspx. Thanks for pointing out the dead link. Upplysning 23:30, 14 November 2014 (CET)
- Thank you, I read it. As a result, I had to do some edits to the page, because the study actually did not find an association between ethnic heterogeneity and life satisfaction, see the updated section. (Galileo 22:39, 15 November 2014 (CET)).
- This should be a working link: https://www.ipsos-mori.com/researchpublications/publications/1226/Life-satisfaction-and-trust-in-other-people.aspx. Thanks for pointing out the dead link. Upplysning 23:30, 14 November 2014 (CET)
- Thank you, I read it. As a result, I had to do some edits to the page, because the study actually did not find an association between ethnic heterogeneity and life satisfaction, see the updated section. (Galileo 22:39, 15 November 2014 (CET)).
- I fail to see support for you claims. The study states, for example, "ethnic diversity in an area does seem to be associated with lower levels of trust and life satisfaction, in line with US evidence".(page 4)
- You stated: "The factor "ethnic heterogeneity" was not found to be statistically significantly associated with life satisfaction using regression analysis". That is misleading. Ethnic heterogeneity/ethnic diversity was not included (at all) in the multiple regression model. But it was examined in the CHAID model which compared the most and least satisfied groups. The study in fact states that "In this study ethnic diversity comes out as the best predictor" of life satisfaction.(page 1)
- So there is certainly enough support to state that there is an association as the study itself in fact does. I have restored the earlier version. Upplysning 07:30, 16 November 2014 (CET)
- Thank you, I read it. As a result, I had to do some edits to the page, because the study actually did not find an association between ethnic heterogeneity and life satisfaction, see the updated section. (Galileo 22:39, 15 November 2014 (CET)).
- I fail to see support for you claims. The study states, for example, "ethnic diversity in an area does seem to be associated with lower levels of trust and life satisfaction, in line with US evidence".(page 4) Upplysning 07:30, 16 November 2014 (CET)
- Yes that is what the report states. Notice the word seem. The report states that ethnic diversity in an area seems to be associated with lower levels of life satisfaction. But on the page you write: "A 2004 study in the UK found that higher ethnic heterogeneity in an area was associated with lower life satisfaction." Why do you omit the word seem? And why do you not mention the next two sentences in the report: "However, it is certain that the relationship is partially with factors associated with higher levels of diversity, such as deprivation, urbanity and population mobility. We need more and better data to unpick the relative importance of these factors." ? (Galileo 20:26, 17 November 2014 (CET))
- I fail to see support for you claims. The study states, for example, "ethnic diversity in an area does seem to be associated with lower levels of trust and life satisfaction, in line with US evidence".(page 4) Upplysning 07:30, 16 November 2014 (CET)
- You stated: "The factor "ethnic heterogeneity" was not found to be statistically significantly associated with life satisfaction using regression analysis". That is misleading. Ethnic heterogeneity/ethnic diversity was not included (at all) in the multiple regression model. Upplysning 07:30, 16 November 2014 (CET)
- It was. The report gives an overview of the relevant variables on page 6 and 7. Ethnic heterogeneity is also included in this list, as you can clearly see. After this overview, we can read (on page 7): "The first stage involves putting all relevant variables into a multiple regression model... ." So it is clear from the report that ethnic heterogeneity was included in the regression analysis. Ethnic heterogeneity was not found to be associated with life satisfaction, because (for example) it is not in the diagram on page 8 of the report (this diagram depicts the results of the multiple regression). (Galileo 20:26, 17 November 2014 (CET)).
- You stated: "The factor "ethnic heterogeneity" was not found to be statistically significantly associated with life satisfaction using regression analysis". That is misleading. Ethnic heterogeneity/ethnic diversity was not included (at all) in the multiple regression model. Upplysning 07:30, 16 November 2014 (CET)
- But it was examined in the CHAID model which compared the most and least satisfied groups. Upplysning 07:30, 16 November 2014 (CET)
- As I explained above, it was also examined in the regression analysis. And the CHAID model not only compared the most and the least satisfied groups, it was also the analysis technique used to identify these groups in the first place. Moreover, the report states on page 10 that we need to be careful when interpreting the outcome of the CHAID analysis, because ethnic heterogeneity is associated with other potentially intervening factors that are not included in the model. In simple words this means that when the analysis is repeated with these other variables included as well, it may well be possible that "ethnic heterogeneity" will not anymore be a group-defining characteristic of the most satisfied group. That’s why the report says we have to be careful interpreting the outcome of the CHAID analysis.(Galileo 20:26, 17 November 2014 (CET)).
- But it was examined in the CHAID model which compared the most and least satisfied groups. Upplysning 07:30, 16 November 2014 (CET)
- The study in fact states that "In this study ethnic diversity comes out as the best predictor" of life satisfaction.(page 1) Upplysning 07:30, 16 November 2014 (CET)
- Only as the best among the area-based factors that were included in this study and as far as I can see from the report only two of these were included: ethnic heterogeneity and the level of deprivation in the area. The best predictors of life satisfaction that come out of the study are in fact views of income and the self-assessed levels of health. (Galileo 20:26, 17 November 2014 (CET)).
- The study in fact states that "In this study ethnic diversity comes out as the best predictor" of life satisfaction.(page 1) Upplysning 07:30, 16 November 2014 (CET)
- So there is certainly enough support to state that there is an association as the study itself in fact does. I have restored the earlier version. Upplysning 07:30, 16 November 2014 (CET)
- As I explained above the study didn't state there is an association. Using regression analysis the study didn't find an association. Only the outcome of the CHAID analysis seemed to show an association, but it is explained in the report that there may be another explanation for this. (Galileo 20:26, 17 November 2014 (CET)).
- So there is certainly enough support to state that there is an association as the study itself in fact does. I have restored the earlier version. Upplysning 07:30, 16 November 2014 (CET)
Do not break up my replies. I have restored my original response.
- Your description is wrong and dishonest as anyone can see who reads the study. I also find it very strange that you so strongly argue against results supporting racial homogeneity. I think that the real reason for your strange antagonism regarding this is that I was involved in getting your crackpot friend Hu1 desysopped for stating among other things that there are 500 millions Jews. I see that you yourself think that this crackpot number is correct [1] and not surprisingly I see that you have been banned from the German Metapedia for crackpottery: [2]. Upplysning 21:22, 17 November 2014 (CET)
- Those other subjects you can discuss with me on the corresponding talk pages. If for example you disagree with me about Einstein's theory you can discuss it with me on the page about his theory (-->Einsteinism). Here on this talk page, I want to discuss the subject "Ethnic heterogeneity". You write that my description is wrong. But your only argument is to say that "anyone who reads the study can see it". In my view, it is actually the opposite. For example, anyone who reads the study can see that "ethnic heterogeneity" was included in the regression analysis, contrary to you claim that it wasn't. (Galileo 21:33, 17 November 2014 (CET))
- All the factors included in the regression analysis is in the image on page 8. Ethnic heterogeneity is not there. Also, obviously the study should have stated explicitly if there was a strong contradiction regarding ethnic heterogeneity between the regression and the CHAID.Upplysning 21:41, 17 November 2014 (CET)
- That is a pertinently untrue statement. For example, do you see any actual measures of income, savings or assets in the image on page 8? No, you can't. But on page 8 we can read: "It is interesting that none of the actual measures of income, savings or assets that were included come out as strong predictors of happiness". This proves that not all factors included in the regression analysis are depicted in the figure on page 8. Only those that had a significant association with life satisfaction are depicted. The percentages give the relative importance of each factor. (Galileo 22:01, 17 November 2014 (CET)).
- It is actually rather unclear what was included in the regression analysis or not. But if ethnic heterogeneity had been included and had not shown the same results as in the CHAID, then this would certainly have been commented on. Note also the statement "We have also matched into the dataset two area-based variables that have had a large impact on other measures of satisfaction in previous studies the level of deprivation in the area (as measured in the Index of Multiple Deprivation) and the level of ethnic heterogeneity." In other words, ethnic heterogeneity was matched for and was therefore not included in the regression.Upplysning 22:11, 17 November 2014 (CET)
- We are talking about ethnic heterogeneity and there is no unclarity about that factor: it was included in the regression analysis, contrary to your claim that it wasn't. It is listed in the list of relevant variables (on page 6 and 7), and the report clearly states that all relevant variables were put in a regression model.
- Regression analysis doesn't always have to show the same result as a CHAID analysis, as you seem to think erroneously. It would be quite easy to construct an example dataset in R to prove this.
- In statistical terminology, the expression "matching for" has a different meaning than "matching into". You confuse these two.
- "Matching into" is what was done in the study reported by Duffy. It basically is a way of adding variables to a dataset. The variables deprivation and ethnic heterogeneity were therefore added to the dataset. "Matching for" means that you are creating two groups (or more) that have for example the same level of ethnic heterogeneity. And that is not was was done in the study. Bobby Duffy correctly used the expression "matched into" and not "matched for".
- So my objection still stands: in the article it is written that "A 2004 study in the UK found that higher ethnic heterogeneity in an area was associated with lower life satisfaction", however, a careful reading of the report of the study reveals that this is a misrepresentation of what was actually reported. (Galileo 23:26, 17 November 2014 (CET)).
- The "list" on pages 6-7 is not a list of the factors in the regression analysis. It is a general list of factors in "the British Household Panel Survey". Exactly what factors are included in the regression in unclear. For example, the image on page 8 includes "Number of GP visits" which is not listed on page 6-7. The study repeatedly states that ethnic heterogeneity/diversity is an important factor and states that the same results have been found in other countries. However, you seem to be completely determined that this should not be the case. Again, this antagonism and determination to state that racial heterogeneity is not harmful is very strange. Upplysning 01:43, 18 November 2014 (CET)
- The list is not exhaustive. This should have been clear to you when reading it. For example, on page 6 you could have read: "Socio-demographic variables, including age, gender, work status, ethnicity, household type, tenure, marital status etc;". So there were othere variables, not included in the list. But one thing is certain: the factors that are listed in the list were certainly included in the regression analysis. Ethnic heterogenetiy was included in the regression analysis.
- Don't falsely accuse me of thinking ethnic heterogeneity is unimportant. I think it is a very important problem. Therefore, I think we should not misrepresent scientific studies like you are doing. That is very harmful to our purpose, because our opponents could easily use it against us and accuse us of "pseudoscience" or something similar. (Galileo 08:26, 18 November 2014 (CET)).
- The "list" on pages 6-7 is not a list of the factors in the regression analysis. It is a general list of factors in "the British Household Panel Survey". Exactly what factors are included in the regression in unclear. For example, the image on page 8 includes "Number of GP visits" which is not listed on page 6-7. The study repeatedly states that ethnic heterogeneity/diversity is an important factor and states that the same results have been found in other countries. However, you seem to be completely determined that this should not be the case. Again, this antagonism and determination to state that racial heterogeneity is not harmful is very strange. Upplysning 01:43, 18 November 2014 (CET)
- It is actually rather unclear what was included in the regression analysis or not. But if ethnic heterogeneity had been included and had not shown the same results as in the CHAID, then this would certainly have been commented on. Note also the statement "We have also matched into the dataset two area-based variables that have had a large impact on other measures of satisfaction in previous studies the level of deprivation in the area (as measured in the Index of Multiple Deprivation) and the level of ethnic heterogeneity." In other words, ethnic heterogeneity was matched for and was therefore not included in the regression.Upplysning 22:11, 17 November 2014 (CET)
- That is a pertinently untrue statement. For example, do you see any actual measures of income, savings or assets in the image on page 8? No, you can't. But on page 8 we can read: "It is interesting that none of the actual measures of income, savings or assets that were included come out as strong predictors of happiness". This proves that not all factors included in the regression analysis are depicted in the figure on page 8. Only those that had a significant association with life satisfaction are depicted. The percentages give the relative importance of each factor. (Galileo 22:01, 17 November 2014 (CET)).
- All the factors included in the regression analysis is in the image on page 8. Ethnic heterogeneity is not there. Also, obviously the study should have stated explicitly if there was a strong contradiction regarding ethnic heterogeneity between the regression and the CHAID.Upplysning 21:41, 17 November 2014 (CET)
- Those other subjects you can discuss with me on the corresponding talk pages. If for example you disagree with me about Einstein's theory you can discuss it with me on the page about his theory (-->Einsteinism). Here on this talk page, I want to discuss the subject "Ethnic heterogeneity". You write that my description is wrong. But your only argument is to say that "anyone who reads the study can see it". In my view, it is actually the opposite. For example, anyone who reads the study can see that "ethnic heterogeneity" was included in the regression analysis, contrary to you claim that it wasn't. (Galileo 21:33, 17 November 2014 (CET))
As noted above the list on pages 6-7 is not a list of the factors in regression. Which factors were included in the regression and/or the CHAID is unclear since this is never exactly listed. What we do know that there is never any statement staging that ethnic heterogeneity/diversity is unimportant. On the contrary. I will give a complete listing of the statements regarding heterogeneity/diversity and life satisfaction in the study:
- "The most satisfied group this analysis could identify are healthy, comfortably-off retired people who live in less ethnically diverse areas and do not want to move. Area-based factors are shown as important throughout the analysis. In this study ethnic diversity comes out as the best predictor, but this will be partly due to the fact that ethnic diversity is also highly related to a number of other area characteristics, such as urbanity and population mobility."(page 1)
- "This study also highlights how important area characteristics are to both levels of happiness and trust in others, and specifically that ethnic diversity in an area does seem to be associated with lower levels of trust and life satisfaction, in line with US evidence 3 . However, it is certain that the relationship is partially with factors associated with higher levels of diversity, such as deprivation, urbanity and population mobility. We need more and better data to unpick the relative importance of these factors."(page 4)
- "We have also matched into the dataset two area-based variables that have had a large impact on other measures of satisfaction in previous studies the level of deprivation in the area (as measured in the Index of Multiple Deprivation) and the level of ethnic heterogeneity. The latter is a measure of the level of ethnic diversity locally that takes account of both the number of different ethnic groups in an area and the proportion in each so, for example, an area with 50% white people and 50% black people will have a lower ethnic heterogeneity score than an area with 50% white, 20% black, 20% Asian and 10% Chinese."(page 7)
- "These are healthy retired people, who do not want to move, but who also see themselves as relatively comfortably off and who live in an area where there is a low level of "ethnic heterogeneity". This is an important factor to come out of the analysis - it is another area-based variable which provides a measure of the ethnic diversity of an area. However, we need to be careful when interpreting this, as ethnic heterogeneity is associated with other potentially intervening factors that are not included in the model - for example, it could be acting as a proxy measure for the urbanity or population churn of areas, given the still very strong concentration of ethnic minority groups in faster-changing urban areas. In any case, it does illustrate again the importance of the nature of area in influencing happiness."(page 10)
Obviously this is a POLITICALLY CORRECT analysis. The rapport wants VERY much to state that ethnic heterogeneity is not negatively related to life satisfaction and SPECULATES regarding this. IF there had in fact been opposing results in the regression, then this would absolutely have been mentioned. See also this: [3] Upplysning 17:19, 18 November 2014 (CET)
- Are you now saying that it is unclear whether ethnic heterogeneity was included in the regression analysis? You now seem to do so. But only two days ago you were claiming that "Ethnic heterogeneity/ethnic diversity was not included (at all) in the multiple regression model.". So your argument doesn't seem to be consistent.
- Anyway, in the report it is mentioned that all relevant factors (i.e. factors that could be expected to have an impact on happiness) were included in a regression model (on page 7 it is written "The first stage involves putting all relevant variables into a multiple regression model"), and ethnic heterogeneity is considered to be relevant in the study (as you seem to admit in your above reply). It is also mentioned in the list of relevant variables on page 6-7.
- Therefore, ethnic heterogeity was included in the regression analysis and the text I inserted ( http://en.metapedia.org/m/index.php?title=Ethnic_heterogeneity&action=historysubmit&diff=381575&oldid=381549 ) was correct. You removed it, and restored your incorrect claim that an association was found in the study ( http://en.metapedia.org/m/index.php?title=Ethnic_heterogeneity&action=historysubmit&diff=381854&oldid=381575 ). But it wasn't. At best, it could be said that there seemed to be an association. The above quotes you give from the report prove me right. None of these quotes support your claim. I will not address your speculation that the report was biased by political correctness, unless you can offer some proof for it. (Galileo 19:05, 18 November 2014 (CET)).
- I have several times states in the past days stated that what is included in the regression analysis is unclear. No, the list on pages 6-7 is not a list of what is in the regression analysis, as you well know. As already states, the regression for example includes "Number of GP visits" but this is not listed on page 6-7. The list on page 6-7 mentions various variables in "the British Household Panel Survey" that may or may not have been used in the regression and/or CHAID and not necessarily in both. Regarding what should be stated in the article, we could use direct quotes instead, likes this: "A 2004 study in the UK stated regarding ethnic heterogeneity that this factor "have had a large impact on other measures of satisfaction in previous studies", "ethnic diversity in an area does seem to be associated with lower levels of trust and life satisfaction, in line with US evidence", and "In this study ethnic diversity comes out as the best predictor, but this will be partly due to the fact that ethnic diversity is also highly related to a number of other area characteristics, such as urbanity and population mobility"". Upplysning 19:50, 18 November 2014 (CET)
- The list on page 6-7 is not an exhaustive list, but it is a list of relevant variables, and the report states that all relevant variables were included in a regression model. This means that all variables in the list were included in a regression model. Since this discussion is about "ethnic heterogeneity" (which is on the list) and since the list isn't exhaustive, your "counterexample" of a variable that was included in the regression analysis but which doesn't appear on the list, proves nothing.
- You claim that there are variables on the list on page 6-7 that were not included in the regression analysis. But since it is stated in the report that "all relevant variables were put in a regression model", your claim implies that some variables in the list were not relevant. That doesn't make sense to me. There is no reason to think that the author of the report would include irrelevant variables in a list and then write above this list (on page 6): "...measures of overall life satisfaction, plus a large number of factors that could be expected to have an impact on happiness. These cover aspects such as:" and then the report gives the list. It is clear that the variables on the list were relevant and that they therefore were included in the regression analysis.
- Regarding your proposal to use direct quotes from the text instead of the current claim that an assocation was found in the 2004 Duffy study: yes, that would be a big improvement of the article, because you go from a claim that is wrong to something that is at least supported by direct quotes. Although, I still think a truthfull and honest representation of the Duffy study demands mention of the outcome of the regression analysis as well, as I did in my edit ( http://en.metapedia.org/m/index.php?title=Ethnic_heterogeneity&action=historysubmit&diff=381575&oldid=381549 ) which you reverted ( http://en.metapedia.org/m/index.php?title=Ethnic_heterogeneity&action=historysubmit&diff=381854&oldid=381575 ). (Galileo 20:22, 18 November 2014 (CET))
- Going around in circles now. See my earlier replies. Again, the list is just a list of potentially useful variables. Exactly which were included in one or both of the two analyses is never clearly stated. Regardless, if the regression has shown politically correct results regarding ethnic heterogeneity, then this would certainly have been VERY prominently mentioned by the very politically correct author who is very anxious to try to explain away ethnic heterogeneity by dubious speculations. Upplysning 20:41, 18 November 2014 (CET)
- And see also my earlier replies to your replies.
- You continue to confuse the issue by saying that "Exactly which were included in one or both of the two analyses is never clearly stated". I never claimed that this was exactly stated, only that the variables listed in the list on page 6-7 were certainly included in the regression analysis and that the list is not exhaustive. I hope you understand the meaning of the term "non-exhaustive list".
- It is quite common, during a CHAID analysis, that an association may seem to exist, but later when more variables are included in the analysis, the association is shown to be non-existent. This is not "dubious speculation" as you write, but a common issue with statistical analyses in general. You are, however, dubiously speculating yourself that the report might have been written by a "politically correct" author. I see no evidence for that speculation. In fact, there is evidence of the opposite: the author found that ethnic heteogeneity was not associated with life satisfaction in a regression analysis, but he didn't mention this in a very prominent way. Clearly an indication of absence of political correctness (in this particular report at least). (Galileo 21:10, 18 November 2014 (CET)).
- Looking at the citations I have given the author is constantly trying to explain away ethnic heterogeneity by speculations. If he had any actual evidence instead of speculations, he would have stated so instead of speculating. Upplysning 21:30, 18 November 2014 (CET)
- If that is true, why are you using such highly speculative texts as sources in Metapedia? Because you can then further add your own speculations, such as your speculation that ethnic heterogeneity was not included in the regression analysis? (Galileo 21:35, 18 November 2014 (CET))
- He is speculating because he does not like the hard facts regarding ethnic heterogeneity and life satisfaction. Similar hard facts have been found in other countries, as the author reluctantly admits. Upplysning 21:38, 18 November 2014 (CET)
- No he is not speculating, he simply points at a possible statistical issue to be resolved in further studies. He openly admits that his results are in line with US evidence and he even gives the reference to it: "Putnam, recent analysis and Participation in heterogeneous communities, Alesina and La Ferrara (2000)". I don't see any "reluctance to admit" anything as you write. This is simply how scientific reports are written, they just state the facts. You cannot expect a scientific report to be written like an article in a tabloid newspaper.(Galileo 21:48, 18 November 2014 (CET))
- Almost every time ethnic heterogeneity is mentioned he at the same time states that the politically incorrect negative relationship with life satisfaction could possibly be due other factors which could maybe be found in other studies - but he never mentions having any evidence for this himself - like from the regression. If he had any, he would mention it. Upplysning 21:55, 18 November 2014 (CET)
- Because the study didn't include potentially intervening factors, the precautions in the report are justified. It is a cardinal sin in statistical research to conclude there is an association, but then later to have to retract this conclusion because the association disappears when adjustments for additional intervening factors are made that were omitted in the first analysis.
- It is good to look for examples of political correctness, but it goes to far (as you do) to see politcal correctness is the scientific study we are discussing right now. Political correctness is that people cannot criticize for example Jews or Negroes. Political correctness is not when a scientific report doesn't put enough emphasis on one of their findings. You are mixing things up. (Galileo 22:24, 18 November 2014 (CET)).
- That ethnic heterogeneity is good is politically correct. The author reportedly tries to explain away the hard fact that it is bad by speculating. If there were any evidence supporting political correctness, it would be VERY prominently mentioned. Upplysning 22:38, 18 November 2014 (CET)
- I don't see how that has anything to do with what we are discussing, because in the study (Duffy, 2004) that we are discussing, NO association (between ethnic heterogeneity and life satisfaction was found using regression analysis. That means not a negative relationship, but also not a positive one. Even if the report would have put a lot of emphasis on that finding, it wouldn't mean it is politically incorrect. And conversely, that the report didn't put a lot of emphasis on this finding cannot be interpreted as political correctness. (Galileo 22:46, 18 November 2014 (CET)).
- Now you are just dishonest, no association in the regression would mean that there were something strange going on with the CHAID. This would support political correctness and prominently mentioned by the very political correct author. Upplysning 22:52, 18 November 2014 (CET)
- I already explained above that the regression analysis doesn't have to show the same result as a CHAID analysis. I even offered to construct an example dataset in R to prove this. So don't accuse me of dishonesty. Good night. (Galileo 22:58, 18 November 2014 (CET)).
- Now you are just dishonest, no association in the regression would mean that there were something strange going on with the CHAID. This would support political correctness and prominently mentioned by the very political correct author. Upplysning 22:52, 18 November 2014 (CET)
- I don't see how that has anything to do with what we are discussing, because in the study (Duffy, 2004) that we are discussing, NO association (between ethnic heterogeneity and life satisfaction was found using regression analysis. That means not a negative relationship, but also not a positive one. Even if the report would have put a lot of emphasis on that finding, it wouldn't mean it is politically incorrect. And conversely, that the report didn't put a lot of emphasis on this finding cannot be interpreted as political correctness. (Galileo 22:46, 18 November 2014 (CET)).
- That ethnic heterogeneity is good is politically correct. The author reportedly tries to explain away the hard fact that it is bad by speculating. If there were any evidence supporting political correctness, it would be VERY prominently mentioned. Upplysning 22:38, 18 November 2014 (CET)
- Almost every time ethnic heterogeneity is mentioned he at the same time states that the politically incorrect negative relationship with life satisfaction could possibly be due other factors which could maybe be found in other studies - but he never mentions having any evidence for this himself - like from the regression. If he had any, he would mention it. Upplysning 21:55, 18 November 2014 (CET)
- No he is not speculating, he simply points at a possible statistical issue to be resolved in further studies. He openly admits that his results are in line with US evidence and he even gives the reference to it: "Putnam, recent analysis and Participation in heterogeneous communities, Alesina and La Ferrara (2000)". I don't see any "reluctance to admit" anything as you write. This is simply how scientific reports are written, they just state the facts. You cannot expect a scientific report to be written like an article in a tabloid newspaper.(Galileo 21:48, 18 November 2014 (CET))
- He is speculating because he does not like the hard facts regarding ethnic heterogeneity and life satisfaction. Similar hard facts have been found in other countries, as the author reluctantly admits. Upplysning 21:38, 18 November 2014 (CET)
- If that is true, why are you using such highly speculative texts as sources in Metapedia? Because you can then further add your own speculations, such as your speculation that ethnic heterogeneity was not included in the regression analysis? (Galileo 21:35, 18 November 2014 (CET))
- Looking at the citations I have given the author is constantly trying to explain away ethnic heterogeneity by speculations. If he had any actual evidence instead of speculations, he would have stated so instead of speculating. Upplysning 21:30, 18 November 2014 (CET)
- Going around in circles now. See my earlier replies. Again, the list is just a list of potentially useful variables. Exactly which were included in one or both of the two analyses is never clearly stated. Regardless, if the regression has shown politically correct results regarding ethnic heterogeneity, then this would certainly have been VERY prominently mentioned by the very politically correct author who is very anxious to try to explain away ethnic heterogeneity by dubious speculations. Upplysning 20:41, 18 November 2014 (CET)
- I have several times states in the past days stated that what is included in the regression analysis is unclear. No, the list on pages 6-7 is not a list of what is in the regression analysis, as you well know. As already states, the regression for example includes "Number of GP visits" but this is not listed on page 6-7. The list on page 6-7 mentions various variables in "the British Household Panel Survey" that may or may not have been used in the regression and/or CHAID and not necessarily in both. Regarding what should be stated in the article, we could use direct quotes instead, likes this: "A 2004 study in the UK stated regarding ethnic heterogeneity that this factor "have had a large impact on other measures of satisfaction in previous studies", "ethnic diversity in an area does seem to be associated with lower levels of trust and life satisfaction, in line with US evidence", and "In this study ethnic diversity comes out as the best predictor, but this will be partly due to the fact that ethnic diversity is also highly related to a number of other area characteristics, such as urbanity and population mobility"". Upplysning 19:50, 18 November 2014 (CET)
Again, it would be mentioned. Let me point out another problem with your claims. You are arguing that those factors not mentioned in the image on page 8 are not associated with life satisfaction. There is no justification for this statement, other factors in the regression may also be associated with life satisfaction although more weakly than those explicitly included in the image. Upplysning 23:10, 18 November 2014 (CET)
- Nice try to prove me wrong. However, on page 7 of the report we can read: "The results from this can be represented in the very simple diagram below. This shows the relative strength of the factors on the right hand side in explaining levels of life satisfaction. The first point to note is that overall the model with these factors included can "explain" 20% of the variation seen in life satisfaction. This is a fairly high proportion for this type of attitudinal variable - but we need to remember that this means that 80% of the variation is unaccounted for - and will be caused by other factors we are not measuring here (such as genes, upbringing and personality).". So this means 20% of the variation in life satisfaction is explained by the variables in the diagram on page 8 (and ethnic heterogeneity is not in the diagram), the other 80% by variables not used in the 2004 Duffy study. 20%+80%=100%. This means that according to the Duffy 2004 report, 100% of the variation in life satisfaction is explained by variables other than ethnic heterogeneity. Ethnic heterogeneity explains 0% of the variation in life satisfaction (according to the regression analysis in Duffy, 2004). Sorry, but these are the facts. By the way, I do not support policies that are aimed to increase ethnic heterogeneity or to force people to live in an existing ethnically heterogenuous area. However, being against such policies does not mean that it is allowed to state false facts. We are always obliged to stay truthful. Also, I think the real problem that is causing the increasing ethnic heterogeneity in European countries is that many people are in fact quite happy about it due to misinformation (think for example about the glorification of many non-white football players by the mass media). Our aim actually should be to explain to people why they should not be happy about increasing ethnic heterogeneity. That would be much better than what you are trying to do (using fallacious arguments, because you cite a study that doesn't support your premise), you are trying to say to the people: don't support increasing ethnic heterogeneity because you are unhappy about it. But it is better to first explain to the people why they should be unhappy about it. In that sense, making happiness itself of central importance like you do could be seen as support of increasing ethnic heterogeneity by those who are happy about it. So in a sense, you are supporting increased ethnic heterogeneity. (Galileo 18:33, 19 November 2014 (CET)).
- I support Upplysning's position regarding this issue. And Galileo, this is the third incident in a short while where your name appears in a context that is contrary to Metapedia's ambitions. I am therefore regretfully forced to revoke your administrator privileges, even though you may still edit and write articles as usual. Aurvandil 23:36, 18 November 2014 (CET)
- Concerning this third incident, I don't see what I did contrary to the ambitions of Metapedia. Please explain it. All I did was checking a reference, and I showed that the reference was misrepresented. (Galileo 18:33, 19 November 2014 (CET)).
- Regarding problematic behaviors, see User_talk:Galileo#Strange_and_damaging_actions. Upplysning 18:37, 19 November 2014 (CET)
- The only thing concerning the present discussion (about ethnic heterogeneity) that you wrote there is: "you [i.e. Galileo] very strangely and fanatically starting demanding that Metapedia should not state that ethnic heterogeneity decreases life satisfaction". But that is a completely false representation of this discussion. It would be perfectly okay if Metapedia would state that ethnic heterogeneity decreases life satisfaction, but not if Metapedia states it while citing a reference that does not support the statement. By using this reference, you are, in a sense, supporting ethnic heterogeneity. (Galileo 18:41, 19 November 2014 (CET)).
- Regarding problematic behaviors, see User_talk:Galileo#Strange_and_damaging_actions. Upplysning 18:37, 19 November 2014 (CET)
- Concerning this third incident, I don't see what I did contrary to the ambitions of Metapedia. Please explain it. All I did was checking a reference, and I showed that the reference was misrepresented. (Galileo 18:33, 19 November 2014 (CET)).
- I support Upplysning's position regarding this issue. And Galileo, this is the third incident in a short while where your name appears in a context that is contrary to Metapedia's ambitions. I am therefore regretfully forced to revoke your administrator privileges, even though you may still edit and write articles as usual. Aurvandil 23:36, 18 November 2014 (CET)