Draft: attached.
Abstract
Criminality rates and fertility varies wildly among Danish immigrant populations by their country of origin. Correlational and regression analyses show that these are very predictable (R's about .85 and .5) at the group level with national IQ, Islam belief, GDP and height as predictors.
Back to [Archive] Post-review discussions
Draft: attached.
Abstract
Criminality rates and fertility varies wildly among Danish immigrant populations by their country of origin. Correlational and regression analyses show that these are very predictable (R's about .85 and .5) at the group level with national IQ, Islam belief, GDP and height as predictors.
Emil, I've already sent you my comments on this paper. To reiterate: 1) the r-value between national criminality and immigrant criminality is very interestingly low; 2) Lynn's 2012 penis size paper is terrible and uses a hoax website for data. Also, "vary widely" would be preferable to "varies wildly" in the abstract. Nevertheless, this paper is interesting and worthy of publication.
Draft: attached.
Abstract
Criminality rates and fertility varies wildly among Danish immigrant populations by their country of origin. Correlational and regression analyses show that these are very predictable (R's about .85 and .5) at the group level with national IQ, Islam belief, GDP and height as predictors.
First, it’d be good if lines were numbered throughout the MS. This should be recommended by admin.
“Luckily, they would.” This style is fine with me but some old fellows (i.e. most academics) may not like it.
Typos: Introduction > line 1: peoeple
Data acquisition> line 3> yieled
“predict of” : page 3. Replace with predictor.
Page 8: replace “analaze” with analyze and dividng with dividing.
Page 9: by by IQ groups. Delete 1 by
“fertility rates for a small numbers of countries is shown” : are shown
Page 9: “A very large Swedish study recently studied the relationship
between height and crime and found it to be negative[25], not positive (except
in their full model, and here it was very slightly positive)”.
Often between groups relationships are different and opposite to those found within populations. Inter group conflicts have a different psychological root than within group violence. It is thus still possible that the positive relationship between height and crime is not only due to the short height of Asians
Perhaps there are some studies done by social psychologists on this issue, it would be worth searching the literature..
Also, another analysis should be run without the East Asian countries, to determine whether Height remains a predictor.
I have fixed a lot of spelling mistakes as well as the other language mistakes pointed out by Duxide above.
I have also run the requested regression without the EAsian countries. The results were surprising. Height was still a predictor without the EAsians and the R was lightly higher. Perhaps height is not merely a proxy but is directly relevant in a cross-population comparison.
---
The r-values between home murder rate and immigrant criminality are very low. In fact, only one of the five r's has a p <.05 and only barely (.048) and the r is measly low at .242. It's probably a fluke. I inserted a little discussion of it.
I have removed the penis length analysis and written a footnote based on your comment.
I have fixed the language error.
I have also fixed a slight data error, where 3 countries had 0 as their GDP value. This seemed not to change results much.
New version attached plus source file.
I have also run the requested regression without the EAsian countries. The results were surprising. Height was still a predictor without the EAsians and the R was lightly higher. Perhaps height is not merely a proxy but is directly relevant in a cross-population comparison.
---
Emil, I've already sent you my comments on this paper. To reiterate: 1) the r-value between national criminality and immigrant criminality is very interestingly low; 2) Lynn's 2012 penis size paper is terrible and uses a hoax website for data. Also, "vary widely" would be preferable to "varies wildly" in the abstract. Nevertheless, this paper is interesting and worthy of publication.
The r-values between home murder rate and immigrant criminality are very low. In fact, only one of the five r's has a p <.05 and only barely (.048) and the r is measly low at .242. It's probably a fluke. I inserted a little discussion of it.
I have removed the penis length analysis and written a footnote based on your comment.
I have fixed the language error.
I have also fixed a slight data error, where 3 countries had 0 as their GDP value. This seemed not to change results much.
New version attached plus source file.
The author has addressed the other reviewer's pointed out problem about dodgy penis size data, by eliminating them. Perhaps in a future publication he can include better penis size data and see whether the model changes significantly. He also established that the correlation between height and crime is not due to East Asians but seems to be a genuine between-groups relationship. A possibility is that this relationship is due to taller groups having a lower SES in Denmark.
Can the author run a separate analysis including ethnic group's SES in Denmark, instead of home country GDP?
Can the author run a separate analysis including ethnic group's SES in Denmark, instead of home country GDP?
The author has addressed the other reviewer's pointed out problem about dodgy penis size data, by eliminating them. Perhaps in a future publication he can include better penis size data and see whether the model changes significantly. He also established that the correlation between height and crime is not due to East Asians but seems to be a genuine between-groups relationship. A possibility is that this relationship is due to taller groups having a lower SES in Denmark.
Can the author run a separate analysis including ethnic group's SES in Denmark, instead of home country GDP?
Unfortunately, the data is not available at this time, as also pointed out in the paper:
This study only analyzed fertility and crime rates because no further data is available for Denmark. However, I have contacted DST and asked if they can release data for immigrant groups by country of origin for other known correlates of IQ, namely educational attainment, income, and being on social benefits. For policy makers, the outcome of future studies examining the predictability of these variables is highly relevant.
I have asked the local stat bureau for the data. It should be forthcoming. No ETA is known.
[quote][/quote]
Unfortunately, the data is not available at this time, as also pointed out in the paper:This study only analyzed fertility and crime rates because no further data is available for Denmark. However, I have contacted DST and asked if they can release data for immigrant groups by country of origin for other known correlates of IQ, namely educational attainment, income, and being on social benefits. For policy makers, the outcome of future studies examining the predictability of these variables is highly relevant.
I have asked the local stat bureau for the data. It should be forthcoming. No ETA is known.
The author has addressed all my criticisms and suggestions. This paper should be published because it presents novel findings on the causes of between-group variability in crime rates.
As per journal requirements.
All the data and exact methods are here:
Main datasheet
Datasheet for 15-19 year olds
Datasheet for 20-29 year olds
Fertility data
Datafile for SPSS attached.
All the data and exact methods are here:
Main datasheet
Datasheet for 15-19 year olds
Datasheet for 20-29 year olds
Fertility data
Datafile for SPSS attached.
I think it should be published now (though possibly the author should try to minimize the use of the first person, which feels jarring to me in an academic journal).
I think it should be published now (though possibly the author should try to minimize the use of the first person, which feels jarring to me in an academic journal).
I think use of the active voice/first person is fine. It's true that it was a long time ago, but I.Newton, C.Darwin and M.Faraday all wrote their essays in the first person. As president of the Royal Society, Lord May said "I would put my own view so strongly as to say that, these days, the use of the passive voice in a research paper is the hallmark of second-rate work. In the long run, more authority is conferred by the direct approach than by the pedantic pretence that some impersonal force is performing the research". May's views were shared by many other eminent scientists, including Astronomer Royal, Martin Rees, who succeded Lord May as president of the Royal Society, and Bruce Alberts, then president of the US National Academy of Sciences.
I personally do not think that the use of the passive voice is the hallmark of second-rate work, but neither is the use of the active voice.
I think it should be published now (though possibly the author should try to minimize the use of the first person, which feels jarring to me in an academic journal).
I (I mean "the author") dislike(s) the use of passive voice which tends to hide the actors. In a multi-author paper, one can also use the active voice to indicate who did what. Instead of the impersonal "A regression analysis was undertaken", one can write [author initials] performed a regression analysis".
The usual reason given for avoiding first person is that reviews are double blind (at least formally) and so one must constantly refer to one's previous papers in the third person.
One thing we wanted to do with this journal is to not focus so much on the form: the formatting, the language, and instead focus on getting science out there freely and quickly. For this reason, we have lax requirements for language and formatting.
The correlation is not a good tool to help you understand the impact of any of your variable in the real world, even if lot of researchers rely on it, sadly enough, and I was also doing the same thing at the very beginning. I have seen it so many times, that a low r or r² were usually associated with meaningful effect, or at least, not small. I have also noticed that many times when I worked on SPSS/EXCEL. I'm not convinced anymore that you should dismiss an effect because it's small only given the standardized coefficient regressions. The odds ratio seems to be a better alternative. Or look for the unstandardized regression coefficient. Sometimes, it's much better. I would say the standardized beta is better to compare the relative impact of your independent var, but not for the "raw" effect.
See here for the odds ratio.
High-Stakes Testing in Higher Education and Employment Appraising the Evidence for Validity and Fairness
EDIT : when I say "unstandardized coefficient" I think you should also be very careful about the scaling of your variable. If, before you performed your regressions, all of your variables have been transformed into z-scores, your unstandardized coefficients won't be of any great use.
See here for the odds ratio.
High-Stakes Testing in Higher Education and Employment Appraising the Evidence for Validity and Fairness
EDIT : when I say "unstandardized coefficient" I think you should also be very careful about the scaling of your variable. If, before you performed your regressions, all of your variables have been transformed into z-scores, your unstandardized coefficients won't be of any great use.
I re-read the article again. Technically, I see no problem. With regard to the data, you have removed the penis length or added a note on it, so no problems remained. Personally, I think it should be published. It's an "ok" for me.
The paper by Emil Kirkegaard, titled "Criminality and fertility among Danish immigrant populations" submitted on 09/03/2014 has been published in Open Differential Psychology, since 3 reviewers have expressed favourable opinion. Post-publication comments are welcome on this forum.
Davide Piffer
Davide Piffer
The final edit before publication.
© Copyright 2024 OpenPsych - All Rights Reserved