"Also can you check the citations to make sure that they are consistent e.g., always before a punctuation mark with a space preceding them."
The citations are consistent. If they occur at the end of a sentence, they are always after the dot, unless they are in a dot-less environment (in a list). There is no dot in they occur in a sentence.
Back to [Archive] Post-review discussions
"Also can you check the citations to make sure that they are consistent e.g., always before a punctuation mark with a space preceding them."
The citations are consistent. If they occur at the end of a sentence, they are always after the dot, unless they are in a dot-less environment (in a list). There is no dot in they occur in a sentence.
Change "it's" to "its", table 2.
4.1: Since they're intercorrelated, why don't you factor analyze those 4 variables? Then you can see if the resulting factor has a stronger correlation to IQ, Age Heaping, Islam, Int.S.Factor and you can perhaps run MCV on it too.
I have completely reworked the paper. It may require some language editing.
New stuff includes weighted analysis, factor analysis (per Piffer's request), and some more stuff from the report.
https://osf.io/e5rxs/
I will add another section re. predictors x crime when controlling for income.
New stuff includes weighted analysis, factor analysis (per Piffer's request), and some more stuff from the report.
https://osf.io/e5rxs/
I will add another section re. predictors x crime when controlling for income.
A new version with the additional section where income is controlled for has been added.
The paper will still need some language editing: https://osf.io/e5rxs/
The paper will still need some language editing: https://osf.io/e5rxs/
I have asked Brian Boutwell to be an external reviewer for this paper. However he has some medical problems. Instead he suggested Michael G. Vaughn (Saint Louis U) and Christopher Salas-Wright (University of Texas) instead. I have asked them if they want to review as well. Hopefully, one of them will review.
Davide Piffer has promised to review it as well. This means that one more reviewer is called for.
Davide Piffer has promised to review it as well. This means that one more reviewer is called for.
A new version with the additional section where income is controlled for has been added.
The paper will still need some language editing: https://osf.io/e5rxs/
I approve publication.
I asked three more biosocial criminologists -- Eric Connolly, Joe Nedelec, Joseph Schwartz -- to review. Two of them agreed to do so. See Twitter: https://twitter.com/KirkegaardEmil/status/634899959272239104
Of course, since authors can only request one external reviewer, if both external reviewers approve, then it only counts as 1.
Of course, since authors can only request one external reviewer, if both external reviewers approve, then it only counts as 1.
Emil,
I only have two minor problems with this paper:
1. On the third page you state:
"National IQ was chosen because cognitive ability is known to cause crime within populations"
Correlation doesn't mean causation. I personally believe that the relationship between the two is largely indirect, i.e., the same selection pressures that have favored cognitive ability have also favored other mental traits: greater capacity for future time orientation, higher thresholds for expression of anger, less monotony avoidance, greater guilt-proneness, greater affective empathy, etc. Pacified societies with weak kinship ties will select for a different mix of mental traits, as Gregory Clark argued in his book. These other traits probably have a greater impact on crime.
In any case, I don't think you wish to argue that increases in cognitive ability cause increases in crime. Yet that is what the above statement says. I would write something more like: "cognitive ability seems to be inversely related to crime rates."
2. On page 9, you state: "If there is any police suspect bias, it will introduce bias in the data." There is a potential for bias in both directions. I would rewrite this statement as follows:
If police target certain communities, i.e., they engage in ethnic/racial profiling, there will be bias in the data. There may also be bias in the other direction. Whether or not "no-go zones" really exist, some immigrant areas have a weaker police presence per capita. Many immigrants are also unwilling to denounce members of their community to the police, either for fear of retaliation or out of ethnic/religious solidarity. This bias would lead to underreporting of crime in such communities.
Other than the above two criticisms, I have no substantial problems with this paper. The following corrections would make the paper more readable.
Page one:
- replace "parted into groups based on their" with "classified by"
- replace "1st to 2nd generation" with "the 1st to 2nd generations"
Page two:
- replace "the data was" with "the data were"
Page three:
- replace "Within countries between individuals" with "Within countries and between individuals"
- replace "predictors is" with "predictors are"
- replace "first generation immigrants" with "first-generation immigrants" (this is a recurring error in the text)
Page four:
-replace "sample immigrants" with "sample of immigrants"
-replace "members of these" with "males of these ages"
-replace "as the request" with "at the request"
-replace "weighing" with "weighting"
-replace "some transformation of it" with "some other mathematical transformation"
- replace "despite controlling for socioeconomic variables" with "even when socioeconomic variables have been controlled"
- replace "welfare receiver" with "welfare recipient"
- replace "All of these" with "All of these reports"
- replace "immigration generation" with "immigrant generation"
Page five:
- replace "sample size for breaking" with "sample sizes to break"
- replace "isn't sufficient" with "is insufficient"
- replace "when controlling for income" with "when income has been controlled"
- replace "Another use of the data in" with "The data may also be used"
- replace "first and second generation" with "the first and second generations"
- replace "2 within generation correlations" with "2 correlations within generations"
- replace "They are shown" with " The results are shown"
- replace ".85 and .83" with "0.85 and 0.83" (this is a recurring error in the text. You should always place a zero before a fractional number)
Page six:
- replace "samples size" with "sample size"
- replace "this indicates" with "there seem to be"
- replace "Crime frequency" with "Crime rate"
- replace "first and second generation" with "the first and second generations"
- replace "Turkish immigrants is" with "Turkish immigrants are"
Page seven:
- replace "Another way to examine it" with "Another way to examine this issue"
Page eight:
- replace "prior to that" with "previously"
- replace "it is not a new phenomenon" with "this overrepresentation is not a new phenomenon"
Page nine:
- replace "causes of crime within each group is" with "causes of crime within each group are"
I only have two minor problems with this paper:
1. On the third page you state:
"National IQ was chosen because cognitive ability is known to cause crime within populations"
Correlation doesn't mean causation. I personally believe that the relationship between the two is largely indirect, i.e., the same selection pressures that have favored cognitive ability have also favored other mental traits: greater capacity for future time orientation, higher thresholds for expression of anger, less monotony avoidance, greater guilt-proneness, greater affective empathy, etc. Pacified societies with weak kinship ties will select for a different mix of mental traits, as Gregory Clark argued in his book. These other traits probably have a greater impact on crime.
In any case, I don't think you wish to argue that increases in cognitive ability cause increases in crime. Yet that is what the above statement says. I would write something more like: "cognitive ability seems to be inversely related to crime rates."
2. On page 9, you state: "If there is any police suspect bias, it will introduce bias in the data." There is a potential for bias in both directions. I would rewrite this statement as follows:
If police target certain communities, i.e., they engage in ethnic/racial profiling, there will be bias in the data. There may also be bias in the other direction. Whether or not "no-go zones" really exist, some immigrant areas have a weaker police presence per capita. Many immigrants are also unwilling to denounce members of their community to the police, either for fear of retaliation or out of ethnic/religious solidarity. This bias would lead to underreporting of crime in such communities.
Other than the above two criticisms, I have no substantial problems with this paper. The following corrections would make the paper more readable.
Page one:
- replace "parted into groups based on their" with "classified by"
- replace "1st to 2nd generation" with "the 1st to 2nd generations"
Page two:
- replace "the data was" with "the data were"
Page three:
- replace "Within countries between individuals" with "Within countries and between individuals"
- replace "predictors is" with "predictors are"
- replace "first generation immigrants" with "first-generation immigrants" (this is a recurring error in the text)
Page four:
-replace "sample immigrants" with "sample of immigrants"
-replace "members of these" with "males of these ages"
-replace "as the request" with "at the request"
-replace "weighing" with "weighting"
-replace "some transformation of it" with "some other mathematical transformation"
- replace "despite controlling for socioeconomic variables" with "even when socioeconomic variables have been controlled"
- replace "welfare receiver" with "welfare recipient"
- replace "All of these" with "All of these reports"
- replace "immigration generation" with "immigrant generation"
Page five:
- replace "sample size for breaking" with "sample sizes to break"
- replace "isn't sufficient" with "is insufficient"
- replace "when controlling for income" with "when income has been controlled"
- replace "Another use of the data in" with "The data may also be used"
- replace "first and second generation" with "the first and second generations"
- replace "2 within generation correlations" with "2 correlations within generations"
- replace "They are shown" with " The results are shown"
- replace ".85 and .83" with "0.85 and 0.83" (this is a recurring error in the text. You should always place a zero before a fractional number)
Page six:
- replace "samples size" with "sample size"
- replace "this indicates" with "there seem to be"
- replace "Crime frequency" with "Crime rate"
- replace "first and second generation" with "the first and second generations"
- replace "Turkish immigrants is" with "Turkish immigrants are"
Page seven:
- replace "Another way to examine it" with "Another way to examine this issue"
Page eight:
- replace "prior to that" with "previously"
- replace "it is not a new phenomenon" with "this overrepresentation is not a new phenomenon"
Page nine:
- replace "causes of crime within each group is" with "causes of crime within each group are"
Thanks Peter. I will look into these matters. I am however going on a trip for the next few days, so I can't make the edits immediately.
Peter,
I have changed the IQ-crime discussion to:
---
I agree with your comments regarding justice system bias. I did not intend to say that it could only be in one direction.
I have rewritten it to:
-
I have made most of the language changes you suggested.
I have not made the change to add a zero before fractional numbers. This practice is unnecessary (there i no extra information by including this) and is often not done in e.g. computer science. I have however added the zeros to the one table that did not have them for consistency.
---
The paper has been updated at OSF.
https://osf.io/e5rxs/
I have changed the IQ-crime discussion to:
National IQ was chosen because cognitive ability is known to be negatively related to crime within populations, with a correlation of about -.2. (9–11). This correlation does not disappear after controlling for socioeconomic status, but it is reduced to about .10 in a within sibling analysis, which may suggest that the crime link to crime is indirect and thru some other trait that covaries with cognitive ability in the population (e.g. self-control, stronger future orientation) but less so within sibling pairs (9). However, some of this reduction in the correlation is probably due to increased measurement error when comparing siblings (9).
---
I agree with your comments regarding justice system bias. I did not intend to say that it could only be in one direction.
I have rewritten it to:
The study relies on suspects of crime, not persons found guilty. If police target certain communities, i.e., they engage in ethnic/racial profiling, there will be bias in the data. It is unclear what the net direction of police bias would be. Members of the police are presumably mostly of Dutch origin, so if they have an outgroup bias, the immigrants most different from the Dutch would receive extra attention. On the other hand, if the police try to avoid naming immigrants as suspects to avoid political outcry, this would result in bias against their own group.
Differential police presence in areas may also cause suspects per capita to be a biased indicator of crime rates. If "no-go zones" exist, some immigrant areas have a weaker police presence per capita which could reduce the suspect rates for persons living in those areas. Many immigrants are also unwilling to denounce members of their community to the police, either for fear of retaliation or out of ethnic/religious solidarity. This bias would lead to underreporting of crime in such communities. For these reasons, would be interesting to examine self-report and verdict based data as has been done for US data (22).
-
I have made most of the language changes you suggested.
I have not made the change to add a zero before fractional numbers. This practice is unnecessary (there i no extra information by including this) and is often not done in e.g. computer science. I have however added the zeros to the one table that did not have them for consistency.
---
The paper has been updated at OSF.
https://osf.io/e5rxs/
I read it once more, and I do not have particular criticisms. So, I won't change my earlier conclusions about approval.
Just two weird sentences :
And one remark, concerning :
Although you have cited these authors, I believe you missed their point. From the above, it is obvious they are talking about what psychometrics is known as measurement invariance. The causes of the differences have different origins. Two groups may thus differ in crime rates due to differences in psychological traits, while the structure of the differences in these traits can differ across groups, which means non-invariance. But the groups may also differ in crime rates due to a mixture of factors (psychological traits + factors unrelated to psychological traits), while the structure of the differences in these factors can differ between groups, which means non-invariance.
Just two weird sentences :
which may suggest that the crime link to crime is indirect
The data may also be used is to examine whether there are differences between first and second-generations in crime rates. I can think of two ways to examine the generational differences.
And one remark, concerning :
“How can we explain the differences? ... These authors state that whoever is interested in the putative causes of ethnic differences should also study the prevalence of risk factors in each ethnic group.” (p. 90) “More insight into the underlying processes and mechanisms among the various groups is required.” (p. 91)
Although you have cited these authors, I believe you missed their point. From the above, it is obvious they are talking about what psychometrics is known as measurement invariance. The causes of the differences have different origins. Two groups may thus differ in crime rates due to differences in psychological traits, while the structure of the differences in these traits can differ across groups, which means non-invariance. But the groups may also differ in crime rates due to a mixture of factors (psychological traits + factors unrelated to psychological traits), while the structure of the differences in these factors can differ between groups, which means non-invariance.
Thanks MH,
I have fixed the two language errors. Files updated.
I don't think the authors know what measurement invariance is (they are not psychometricians), so it seems unlikely that they are referring to that, at least directly. But you are right that one could view this as a measurement invariance discussion. Let's say two groups, A and B, differ in come crime outcome Y. If MI holds, the relationship of C to causes X, Y, and Z should be similar for the two groups, i.e. similar intercepts, slope and correlation. Furthermore, the difference between the means of Y should be due to differences in X, Y and Z between the groups.
My working hypothesis is that the causes of crime between the groups are similar, but that the groups differ in their mean levels of these causes (various psychological traits and cultural beliefs; in this paper focusing on cognitive ability and Muslim beliefs) which gives rise to the observed differences.
I have fixed the two language errors. Files updated.
I don't think the authors know what measurement invariance is (they are not psychometricians), so it seems unlikely that they are referring to that, at least directly. But you are right that one could view this as a measurement invariance discussion. Let's say two groups, A and B, differ in come crime outcome Y. If MI holds, the relationship of C to causes X, Y, and Z should be similar for the two groups, i.e. similar intercepts, slope and correlation. Furthermore, the difference between the means of Y should be due to differences in X, Y and Z between the groups.
My working hypothesis is that the causes of crime between the groups are similar, but that the groups differ in their mean levels of these causes (various psychological traits and cultural beliefs; in this paper focusing on cognitive ability and Muslim beliefs) which gives rise to the observed differences.
They can understand the concept of measurement invariance without knowing how it's called. I believe they understand this because they said explicitly "Literature on ethnic differences in crime tends to focus more attention on the prevalence (the extent of occurrence) of ethnic differences in criminal behaviour than on the question whether the causes of such behaviour are the same among the various ethnic groups (Loeber and Farrington, 2004)." that it is not the overall difference we should investigate but the structure of this difference.
"[...] which may suggest that the crime link to crime is indirect and thru some other trait [...]"
I would write:
"The link to crime may therefore be indirect and act through some other trait [...]"
"The study relies on suspects of crime, not persons found guilty."
I would write:
"The study relies on crime suspects, not on persons found guilty."
"If "no-go zones" exist, some immigrant areas have a weaker police presence per capita which could reduce the suspect rates for persons living in those areas."
A "no-go zone" is not an area with a weaker police presence. It's an area where the police simply do not go. The term itself is controversial and should perhaps be avoided altogether.
"For these reasons, would be interesting to examine self-report and verdict based data as has been done for US data (22)."
You should insert "it" before "would."
I would write:
"The link to crime may therefore be indirect and act through some other trait [...]"
"The study relies on suspects of crime, not persons found guilty."
I would write:
"The study relies on crime suspects, not on persons found guilty."
"If "no-go zones" exist, some immigrant areas have a weaker police presence per capita which could reduce the suspect rates for persons living in those areas."
A "no-go zone" is not an area with a weaker police presence. It's an area where the police simply do not go. The term itself is controversial and should perhaps be avoided altogether.
"For these reasons, would be interesting to examine self-report and verdict based data as has been done for US data (22)."
You should insert "it" before "would."
Peter,
Thanks for additional comments re. language.
This one was already changed in the latest version:
This correlation does not disappear after controlling for socioeconomic status, but it is reduced to about .10 in a within sibling analysis, which may suggest that the link to crime is indirect and thru some other trait that covaries with cognitive ability in the population (e.g. self-control, stronger future orientation) but less so within sibling pairs (9).
(Yes, I wrote "thru" on purpose. I'm a spelling reform advocate and try to smuggle them in when possible!)
I used your version.
In Denmark, we have several areas where police generally do not go unless by force (i.e. multiple units). Does this count? Sweden has lots of these areas. There was some attention to this recently: http://swedenreport.org/2014/10/29/swedish-police-55-official-no-go-zones/ I did a quick search, but didn't find any reliable information re. immigrant area areas for the Netherlands.
I changed the sentence to: Some immigrant areas probably have a weaker police presence per capita which could reduce the suspect rates for persons living in those areas.
Fixed.
Files updated.
Thanks for additional comments re. language.
"[...] which may suggest that the crime link to crime is indirect and thru some other trait [...]"
I would write:
"The link to crime may therefore be indirect and act through some other trait [...]"
This one was already changed in the latest version:
This correlation does not disappear after controlling for socioeconomic status, but it is reduced to about .10 in a within sibling analysis, which may suggest that the link to crime is indirect and thru some other trait that covaries with cognitive ability in the population (e.g. self-control, stronger future orientation) but less so within sibling pairs (9).
(Yes, I wrote "thru" on purpose. I'm a spelling reform advocate and try to smuggle them in when possible!)
"The study relies on suspects of crime, not persons found guilty."
I would write:
"The study relies on crime suspects, not on persons found guilty."
I used your version.
"If "no-go zones" exist, some immigrant areas have a weaker police presence per capita which could reduce the suspect rates for persons living in those areas."
A "no-go zone" is not an area with a weaker police presence. It's an area where the police simply do not go. The term itself is controversial and should perhaps be avoided altogether.
In Denmark, we have several areas where police generally do not go unless by force (i.e. multiple units). Does this count? Sweden has lots of these areas. There was some attention to this recently: http://swedenreport.org/2014/10/29/swedish-police-55-official-no-go-zones/ I did a quick search, but didn't find any reliable information re. immigrant area areas for the Netherlands.
I changed the sentence to: Some immigrant areas probably have a weaker police presence per capita which could reduce the suspect rates for persons living in those areas.
"For these reasons, would be interesting to examine self-report and verdict based data as has been done for US data (22)."
You should insert "it" before "would."
Fixed.
Files updated.
Since apparently no review is forthcoming from the criminologists I tried, I tried 4 more:
Anthony Walsh
Professor, Boise State University
http://sspa.boisestate.edu/criminaljustice/faculty-staff/anthony-walsh-ph-d/
twalsh@boisestate.edu
Kevin M. Beaver
Professor,
http://criminology.fsu.edu/faculty-and-staff/college-faculty/kevin-beaver/
kbeaver@fsu.edu
John Paul Wright
Professor, University of Cincinnati
http://cech.uc.edu/criminaljustice/employees.html?eid=wrightww
john.wright@uc.edu
J C Barnes
Associate Professor
http://cech.uc.edu/criminaljustice/employees.html?eid=barnejr
jc.barnes@uc.edu
This was 2 days ago. If they don't respond either, I will try to get Meisenberg.
Anthony Walsh
Professor, Boise State University
http://sspa.boisestate.edu/criminaljustice/faculty-staff/anthony-walsh-ph-d/
twalsh@boisestate.edu
Kevin M. Beaver
Professor,
http://criminology.fsu.edu/faculty-and-staff/college-faculty/kevin-beaver/
kbeaver@fsu.edu
John Paul Wright
Professor, University of Cincinnati
http://cech.uc.edu/criminaljustice/employees.html?eid=wrightww
john.wright@uc.edu
J C Barnes
Associate Professor
http://cech.uc.edu/criminaljustice/employees.html?eid=barnejr
jc.barnes@uc.edu
This was 2 days ago. If they don't respond either, I will try to get Meisenberg.
Instead, I wrote to Curtis Dunkel.
Associate Professor
Western Illinois University, Macomb · Psychology
http://www.wiu.edu/cas/psychology/faculty/dunkelcurtis/
https://www.researchgate.net/profile/Curtis_Dunkel/publications
Who accepted to review.
Associate Professor
Western Illinois University, Macomb · Psychology
http://www.wiu.edu/cas/psychology/faculty/dunkelcurtis/
https://www.researchgate.net/profile/Curtis_Dunkel/publications
Who accepted to review.
Instead, I wrote to Curtis Dunkel.
Associate Professor
Western Illinois University, Macomb · Psychology
http://www.wiu.edu/cas/psychology/faculty/dunkelcurtis/
https://www.researchgate.net/profile/Curtis_Dunkel/publications
Who accepted to review.
I hope the review is helpful. The primary suggestion I have is to supply more background (i.e., greater detail concerning your previous papers). As it reads now, there seems to be an assumption of familiarity with the previous work. I think the paper should be able to stand alone.
Curt,
Thanks for reviewing. I will reply to the points on the forum including quotes so others who don't want to open the PDF can see what I am replying to.
(This refers to the abstract.)
The numbers given in the abstract are a mean over the two age groups. It is calculated from Table 1, 12_17men and 18_24men. For the generations: 12_17_1gen, 18_24_1gen, 12_17_2gen, and 18_24_2gen.
I have reworded it to:
The crime rates were correlated with country of origin predictor variables: national IQ, prevalence of Islam and general socioeconomic factor (S). For males aged 12-17 and 18-24, the mean correlation with IQ, Islam, and S was, respectively, -.51, .37, and -.42. When subsamples split into 1st and 2nd generations were used, the mean correlation was -.74, .34, and -.40. A general crime factor among young persons was extracted. The correlations with the predictors for this variable were -.80, .34, and -.43. The results were similar when weighing the observations by the population of each immigrant group in the Netherlands. The results were also similar when using crime rates controlled for differences in household income.
Is this ok?
--
You have crossed out "to my knowledge" on p. 1. I want to keep this because I want to leave open the possibility that someone has examined the question without me being aware of it.
--
Pun noted. ;) But all the "generals" serve a purpose.
1) Without this qualifier the claim could be read as a strict claim about the immigrants retaining their e.g. cognitive ability levels. I want to leave open the option for gains, especially for immigrants from very deprived circumstances such as severe malnutrition.
2+3) The hypothesis is general in the sense that it is not specific. Many explanations offered for socioeconomic outcomes are too specific and miss the more general pattern. Plausibly (Occam's Razor), a general pattern requires a general hypotheses, not a number of independent hypotheses that also jointly explain the phenomenon. It also connects with the discussion with the quote from the Dutch report.
--
The individual wants to remain anonymous I think. It is not worth doing a fake quotation along the lines of (Anonymous, personal communication, 2014).
The previous papers are cited just above, references 1-4. But I can add them again here if you wish. The first paragraph of the introduction already introduces the topic.
In general, what we have is a disagreement re. how much background knowledge is assumed. I am writing to a fairly specialist audience and so I don't want to go into details, readers can either read the citations or already know them.
More importantly, because I am doing so many papers, it takes too much time to explain the same thing over and over again, so I like to keep it more short and to the point. This helps the busy reader because he does not have to read introductions to the same topic over and over in 20 papers. I usually skip overly long introductions when reading papers.
Nevertheless, I have changed it to:
I was contacted by a Dutch researcher who had read my previous papers on the performance of immigrants (1–4).
Is this fine with you?
--
It is described in the references given. However, I have added the following footnote:
“Among sociologists, in particular, there is a tendency to interpret the correlation between a social variable and phenotype as a causal relation, without even considering the possibility that genetic influences might be behind the correlation, making it completely bogus.” (16). Controlling for income/social status when comparing crime rates assumes that income/social status is a purely environmental variable not caused by immigrant characteristics, whereas in all likelihood immigrant characteristics cause both crime and income.
A fallacy is an error is reasoning. In this cause it is an error in interpreting a path and what happens when it is controlled for. This is similar usage to other errors of reasoning such as gambler's fallacy (failure to realize that e.g. dice throws are statistically independent events meaning that past throws provide no guide to future throws) or other causal reasoning errors (post hoc ergo propter hoc, cum hoc ergo propter hoc). My usage in is line with common use in logic and critical thinking as well as the (limited) psychological literature.
Is the change ok with you?
--
Fixed.
--
Yes, but controlling is expected to cause a decrease in the size of the differences. Not because income is thought to be causal (it may be), but because of the way SEM works. To be brief, the network we have is:
[attachment=680]
(Plot drawn after Beaujean's Latent Variable Modeling Using R, 2014, p. 57.)
--
Presented in the next two subsections (3.1, 3.2). Is this unclear? It seems redundant to mention them.
--
It is copied from a Danish report, except that I have exchanged the Danish text with added English translations.
The numbers given in this report are not absolute crime rates, but relative crime rates, hence not the same as those in the picture. Furthermore, they are relative crime suspect rates. The Danish numbers concern persons found guilty per capita.
I would rather not have to make a figure like this. One of my points with the series of articles into immigrant groups is that one should not just rely on these macro-origin classifications since they hide crucial differences. E.g. for Denmark, Chinese are included in the non-Western group, but Chinese are actually about 1/3 as crime prone as Danish natives. I see no reason to group them with e.g. Turks.
--
Sure, e.g. the immigrant report cited previously. Or you could get the numbers from the database yourself. Everything is available in English.
http://www.statistikbanken.dk/statbank5a/SelectVarVal/Define.asp?MainTable=FOLK2&PLanguage=1&PXSId=0&wsid=cftree
As of 2015, 61634 Turks in Denmark. Total population is 5659715. Thus they are about 1.1%. In Denmark, immigrants are very diverse (many origin countries). They form about 10% of all immigrants in Denmark.
I have added a footnote with:
Turks constitute about 1.1% of the total population and about 10% of the total immigrant population. The total immigrant population constitute about 11.6% of the total population of Denmark, but see (19,20).
Is this ok? The specific source is already cited in the table caption (FOLK2, which is what the link above goes to).
--
The reference above concerns crime in homelands, not of immigrants in host nations which is what the present study is concerned with. Note that one prior study did look at home country murder rate x crime rate in DK, but found no relationship. Surprising.
http://openpsych.net/ODP/2014/03/criminality-and-fertility-among-danish-immigrant-populations/
--
I have updated the files at OSF with the new changes.
https://osf.io/e5rxs/
Thanks for reviewing. I will reply to the points on the forum including quotes so others who don't want to open the PDF can see what I am replying to.
(This refers to the abstract.)
Should read "for both groups of males between 12-17 and 18-24" because in the analysis you keep (correctly I believe) those age groups separate.
The numbers given in the abstract are a mean over the two age groups. It is calculated from Table 1, 12_17men and 18_24men. For the generations: 12_17_1gen, 18_24_1gen, 12_17_2gen, and 18_24_2gen.
I have reworded it to:
The crime rates were correlated with country of origin predictor variables: national IQ, prevalence of Islam and general socioeconomic factor (S). For males aged 12-17 and 18-24, the mean correlation with IQ, Islam, and S was, respectively, -.51, .37, and -.42. When subsamples split into 1st and 2nd generations were used, the mean correlation was -.74, .34, and -.40. A general crime factor among young persons was extracted. The correlations with the predictors for this variable were -.80, .34, and -.43. The results were similar when weighing the observations by the population of each immigrant group in the Netherlands. The results were also similar when using crime rates controlled for differences in household income.
Is this ok?
--
You have crossed out "to my knowledge" on p. 1. I want to keep this because I want to leave open the possibility that someone has examined the question without me being aware of it.
--
To many generals in this paragraph.
Pun noted. ;) But all the "generals" serve a purpose.
1) Without this qualifier the claim could be read as a strict claim about the immigrants retaining their e.g. cognitive ability levels. I want to leave open the option for gains, especially for immigrants from very deprived circumstances such as severe malnutrition.
2+3) The hypothesis is general in the sense that it is not specific. Many explanations offered for socioeconomic outcomes are too specific and miss the more general pattern. Plausibly (Occam's Razor), a general pattern requires a general hypotheses, not a number of independent hypotheses that also jointly explain the phenomenon. It also connects with the discussion with the quote from the Dutch report.
--
Put in reference and describe the results of the previous papers. As it is there is an assumption that the reader has read the other papers.
The individual wants to remain anonymous I think. It is not worth doing a fake quotation along the lines of (Anonymous, personal communication, 2014).
The previous papers are cited just above, references 1-4. But I can add them again here if you wish. The first paragraph of the introduction already introduces the topic.
In general, what we have is a disagreement re. how much background knowledge is assumed. I am writing to a fairly specialist audience and so I don't want to go into details, readers can either read the citations or already know them.
More importantly, because I am doing so many papers, it takes too much time to explain the same thing over and over again, so I like to keep it more short and to the point. This helps the busy reader because he does not have to read introductions to the same topic over and over in 20 papers. I usually skip overly long introductions when reading papers.
Nevertheless, I have changed it to:
I was contacted by a Dutch researcher who had read my previous papers on the performance of immigrants (1–4).
Is this fine with you?
--
What is the sociologist's fallacy? Fallacy seems like an awfully strong word to use. Take a sentence or two to explain.
It is described in the references given. However, I have added the following footnote:
“Among sociologists, in particular, there is a tendency to interpret the correlation between a social variable and phenotype as a causal relation, without even considering the possibility that genetic influences might be behind the correlation, making it completely bogus.” (16). Controlling for income/social status when comparing crime rates assumes that income/social status is a purely environmental variable not caused by immigrant characteristics, whereas in all likelihood immigrant characteristics cause both crime and income.
A fallacy is an error is reasoning. In this cause it is an error in interpreting a path and what happens when it is controlled for. This is similar usage to other errors of reasoning such as gambler's fallacy (failure to realize that e.g. dice throws are statistically independent events meaning that past throws provide no guide to future throws) or other causal reasoning errors (post hoc ergo propter hoc, cum hoc ergo propter hoc). My usage in is line with common use in logic and critical thinking as well as the (limited) psychological literature.
Is the change ok with you?
--
substitute "results" for "conclusions"
Fixed.
--
"as expected", but isn't that the sociologists fallacy?
Yes, but controlling is expected to cause a decrease in the size of the differences. Not because income is thought to be causal (it may be), but because of the way SEM works. To be brief, the network we have is:
[attachment=680]
(Plot drawn after Beaujean's Latent Variable Modeling Using R, 2014, p. 57.)
--
The two ways are...
Presented in the next two subsections (3.1, 3.2). Is this unclear? It seems redundant to mention them.
--
I really like this Figure. Could you categorize western and non-western in the current study and do essentially the same thing?
It is copied from a Danish report, except that I have exchanged the Danish text with added English translations.
The numbers given in this report are not absolute crime rates, but relative crime rates, hence not the same as those in the picture. Furthermore, they are relative crime suspect rates. The Danish numbers concern persons found guilty per capita.
I would rather not have to make a figure like this. One of my points with the series of articles into immigrant groups is that one should not just rely on these macro-origin classifications since they hide crucial differences. E.g. for Denmark, Chinese are included in the non-Western group, but Chinese are actually about 1/3 as crime prone as Danish natives. I see no reason to group them with e.g. Turks.
--
Is a citation available? What is the percentage or raw number?
Sure, e.g. the immigrant report cited previously. Or you could get the numbers from the database yourself. Everything is available in English.
http://www.statistikbanken.dk/statbank5a/SelectVarVal/Define.asp?MainTable=FOLK2&PLanguage=1&PXSId=0&wsid=cftree
As of 2015, 61634 Turks in Denmark. Total population is 5659715. Thus they are about 1.1%. In Denmark, immigrants are very diverse (many origin countries). They form about 10% of all immigrants in Denmark.
I have added a footnote with:
Turks constitute about 1.1% of the total population and about 10% of the total immigrant population. The total immigrant population constitute about 11.6% of the total population of Denmark, but see (19,20).
Is this ok? The specific source is already cited in the table caption (FOLK2, which is what the link above goes to).
--
I know there is a lot out there from Kevin Beaver, but this reference popped in my head. I think it is worth including.
Rushton, J. P., & Templer, D. I. (2009). National differences in intelligence, crime, income, and skin color. Intelligence, 37, 341-346.
The reference above concerns crime in homelands, not of immigrants in host nations which is what the present study is concerned with. Note that one prior study did look at home country murder rate x crime rate in DK, but found no relationship. Surprising.
http://openpsych.net/ODP/2014/03/criminality-and-fertility-among-danish-immigrant-populations/
--
I have updated the files at OSF with the new changes.
https://osf.io/e5rxs/
Curt,
Thanks for reviewing. I will reply to the points on the forum including quotes so others who don't want to open the PDF can see what I am replying to.
(This refers to the abstract.)Should read "for both groups of males between 12-17 and 18-24" because in the analysis you keep (correctly I believe) those age groups separate.
The numbers given in the abstract are a mean over the two age groups. It is calculated from Table 1, 12_17men and 18_24men. For the generations: 12_17_1gen, 18_24_1gen, 12_17_2gen, and 18_24_2gen.
I have reworded it to:
The crime rates were correlated with country of origin predictor variables: national IQ, prevalence of Islam and general socioeconomic factor (S). For males aged 12-17 and 18-24, the mean correlation with IQ, Islam, and S was, respectively, -.51, .37, and -.42. When subsamples split into 1st and 2nd generations were used, the mean correlation was -.74, .34, and -.40. A general crime factor among young persons was extracted. The correlations with the predictors for this variable were -.80, .34, and -.43. The results were similar when weighing the observations by the population of each immigrant group in the Netherlands. The results were also similar when using crime rates controlled for differences in household income.
Is this ok?
--
You have crossed out "to my knowledge" on p. 1. I want to keep this because I want to leave open the possibility that someone has examined the question without me being aware of it.
--To many generals in this paragraph.
Pun noted. ;) But all the "generals" serve a purpose.
1) Without this qualifier the claim could be read as a strict claim about the immigrants retaining their e.g. cognitive ability levels. I want to leave open the option for gains, especially for immigrants from very deprived circumstances such as severe malnutrition.
2+3) The hypothesis is general in the sense that it is not specific. Many explanations offered for socioeconomic outcomes are too specific and miss the more general pattern. Plausibly (Occam's Razor), a general pattern requires a general hypotheses, not a number of independent hypotheses that also jointly explain the phenomenon. It also connects with the discussion with the quote from the Dutch report.
--Put in reference and describe the results of the previous papers. As it is there is an assumption that the reader has read the other papers.
The individual wants to remain anonymous I think. It is not worth doing a fake quotation along the lines of (Anonymous, personal communication, 2014).
The previous papers are cited just above, references 1-4. But I can add them again here if you wish. The first paragraph of the introduction already introduces the topic.
In general, what we have is a disagreement re. how much background knowledge is assumed. I am writing to a fairly specialist audience and so I don't want to go into details, readers can either read the citations or already know them.
More importantly, because I am doing so many papers, it takes too much time to explain the same thing over and over again, so I like to keep it more short and to the point. This helps the busy reader because he does not have to read introductions to the same topic over and over in 20 papers. I usually skip overly long introductions when reading papers.
Nevertheless, I have changed it to:
I was contacted by a Dutch researcher who had read my previous papers on the performance of immigrants (1–4).
Is this fine with you?
--What is the sociologist's fallacy? Fallacy seems like an awfully strong word to use. Take a sentence or two to explain.
It is described in the references given. However, I have added the following footnote:
“Among sociologists, in particular, there is a tendency to interpret the correlation between a social variable and phenotype as a causal relation, without even considering the possibility that genetic influences might be behind the correlation, making it completely bogus.” (16). Controlling for income/social status when comparing crime rates assumes that income/social status is a purely environmental variable not caused by immigrant characteristics, whereas in all likelihood immigrant characteristics cause both crime and income.
A fallacy is an error is reasoning. In this cause it is an error in interpreting a path and what happens when it is controlled for. This is similar usage to other errors of reasoning such as gambler's fallacy (failure to realize that e.g. dice throws are statistically independent events meaning that past throws provide no guide to future throws) or other causal reasoning errors (post hoc ergo propter hoc, cum hoc ergo propter hoc). My usage in is line with common use in logic and critical thinking as well as the (limited) psychological literature.
Is the change ok with you?
--substitute "results" for "conclusions"
Fixed.
--"as expected", but isn't that the sociologists fallacy?
Yes, but controlling is expected to cause a decrease in the size of the differences. Not because income is thought to be causal (it may be), but because of the way SEM works. To be brief, the network we have is:
(Plot drawn after Beaujean's Latent Variable Modeling Using R, 2014, p. 57.)
--The two ways are...
Presented in the next two subsections (3.1, 3.2). Is this unclear? It seems redundant to mention them.
--I really like this Figure. Could you categorize western and non-western in the current study and do essentially the same thing?
It is copied from a Danish report, except that I have exchanged the Danish text with added English translations.
The numbers given in this report are not absolute crime rates, but relative crime rates, hence not the same as those in the picture. Furthermore, they are relative crime suspect rates. The Danish numbers concern persons found guilty per capita.
I would rather not have to make a figure like this. One of my points with the series of articles into immigrant groups is that one should not just rely on these macro-origin classifications since they hide crucial differences. E.g. for Denmark, Chinese are included in the non-Western group, but Chinese are actually about 1/3 as crime prone as Danish natives. I see no reason to group them with e.g. Turks.
--Is a citation available? What is the percentage or raw number?
Sure, e.g. the immigrant report cited previously. Or you could get the numbers from the database yourself. Everything is available in English.
http://www.statistikbanken.dk/statbank5a/SelectVarVal/Define.asp?MainTable=FOLK2&PLanguage=1&PXSId=0&wsid=cftree
As of 2015, 61634 Turks in Denmark. Total population is 5659715. Thus they are about 1.1%. In Denmark, immigrants are very diverse (many origin countries). They form about 10% of all immigrants in Denmark.
I have added a footnote with:
Turks constitute about 1.1% of the total population and about 10% of the total immigrant population. The total immigrant population constitute about 11.6% of the total population of Denmark, but see (19,20).
Is this ok? The specific source is already cited in the table caption (FOLK2, which is what the link above goes to).
--I know there is a lot out there from Kevin Beaver, but this reference popped in my head. I think it is worth including.
Rushton, J. P., & Templer, D. I. (2009). National differences in intelligence, crime, income, and skin color. Intelligence, 37, 341-346.
The reference above concerns crime in homelands, not of immigrants in host nations which is what the present study is concerned with. Note that one prior study did look at home country murder rate x crime rate in DK, but found no relationship. Surprising.
http://openpsych.net/ODP/2014/03/criminality-and-fertility-among-danish-immigrant-populations/
--
I have updated the files at OSF with the new changes.
https://osf.io/e5rxs/
Emil, I read through the manuscript again. I still think a few minor changes will make it easier for readers to follow. I tried to be more specific in my suggestions. Note, these are differences of opinion concerning style and not content. Thank you for the opportunity to review the paper. Curt
© Copyright 2024 OpenPsych - All Rights Reserved