Talk:Immigration and crime

Wiki Education assignment: Adding Immigrants Quantitative Sources for Latinx Immigration History

This article was the subject of a Wiki Education Foundation-supported course assignment, between 17 January 2022 and 6 May 2022. Further details are available on the course page. Student editor(s): Estanley25 (article contribs).

Primary sources and synthesis

@Zilch-nada, the article before my bold edits that you have attempted to revert, was using lots of improper WP:synthesis of primary sources to make broad arguments.
Instead, I replaced (or flagged) those primary sources (which were not necessarily reliable sources that were also not as current as would be ideal) with secondary WP:Reliable sources that make more global claims and are more WP:verifiable.
I am really confused as to what the disagreement is about and extremely confident in the edits I made. Can you be more specific as to what your issues are with these edits? Superb Owl (talk) 19:22, 5 August 2024 (UTC)

For instance, the source of Cormier states that immigrants are less like to commit crimes in just about every country. Whereas, we have country-by-country sources showing immigrants committing more crime. It is not SYNTHESIS to doubt the former; there is an outright contradiction in sources.

A says: Just about no countries.

B, C, D, and so on: examples of countries that do - that is a direct, not indirect, negation.

My reverting was because you substantially changed every and all substance of the notion of a mixed relationship, and replaced them with describing them as having no relationship. That is an incredibly substantial different article, you must admit?

I am doubtful of most of the current wording. I think at least Cormier's source should be removed, at least in lede via WEIGHT. Zilch-nada (talk) 12:52, 6 August 2024 (UTC)

@Zilch-nada, which countries have reliable secondary sources (such as meta-analyses) saying that immigration increases crime? When I found this article it cited maybe 2 primary sources from 2 different countries from 20 years ago that showed some correlation, not causation. That certainly should not be in the lead, and certainly not synthesized to make any broader claims outside of those studies.
Cormier is a secondary source. I agree it would be helpful to see what he is referencing more specifically and will add that. I can keep finding more secondary sources, but have not found one so far that claims a significant link between the two despite the suggestion made by this article or specific sections in it (e.g. Sweden, Denmark and waiting for verification of Finland, the other country cited in the former lead that was being used to make a claim that so far has not held up well). Superb Owl (talk) 16:08, 6 August 2024 (UTC)

The individual sources of countries, such as Sweden, etc., indicate higher rates of crime. Now, I understand that it is wrong to extrapolate - synthesise - that Europe as a whole is like that - or indeed any causation; of course I understand that. But this article puts particular and excessive emphasis, for example, on the United States in the lede. There are nearly 200 countries, and 10s if not 100s of millions of immigrants globally. Why then does this article emphasise in the lede that the United States of all countries shows no relationship? So my idea would be not to refer to any one country in particular unless other countries are mentioned. It seems very US-centric just to include the US. Zilch-nada (talk) 17:31, 8 August 2024 (UTC)

The U.S. has the most reliable data and so is given more prominence in the lead. If you have great meta-analyses and secondary sources of Europe (similar population size to US) then that would deserve equal weight in my opinion.

You still need to prove that Sweden indicates higher rates of crime since the only secondary source I have found says it does not. I am reverting the last 4 edits, which all appear POV. If you have reliable sources that I am not seeing, then yes, we can make some of the changes that have been made but otherwise it is not supported by the evidence. Superb Owl (talk) 17:53, 8 August 2024 (UTC)

You clearly haven't seen my edits then. Some were simple reordering. Zilch-nada (talk) 17:54, 8 August 2024 (UTC)

The first sentence/paragraph of a long section like Sweden was a good recent reliable summary of the data. That should go first.
I have not had time to sift through all the sources below it, but the ones I have made it to have not all been reliable or reliably summarized. I can do that now as it clearly needs some work as has much of this article Superb Owl (talk) 17:58, 8 August 2024 (UTC)

No, Lindberg describes the general discourse and states that it's wrong (without much elaboration), whereas the next sentences describes the actual meat of the article; "Those with immigrant background are over-represented in Swedish crime statistics". That's much more of a summary. Zilch-nada (talk) 18:00, 8 August 2024 (UTC)

Uh... Those with immigrant background are over-represented in Swedish crime statistics. Research shows that socioeconomic factors, such as unemployment, poverty, exclusion language, and other skills explain most of difference in crime rates between immigrants and natives.

Immigrants are over-represented, as described in these sources = higher rates of crime among immigrants. Also, that is an important distinction. We need to distinguish between immigration's effect on the overall crime rate with the rates of crime associated with immigrants. Both evidently appear to be relevant. Zilch-nada (talk) 17:58, 8 August 2024 (UTC)

Correlation does not equal causation. Causation studies are more relevant than correlation studies. They should have priority, especially when correlation is often used to imply causation Superb Owl (talk) 18:01, 8 August 2024 (UTC)

Causation of immigrants committing more crime is literally explained there? You seem to be focused on immigration's effect on the overall crime rate (say causation A) as opposed to rates of crime associated with immigrants (say causation B). Causation B is quite literally what I just quoted. This article is more than merely about (1.) immigration's effect on the overall crime rate, but also (2.) the crime rate among immigrants. Those are two relevant pieces of info. Zilch-nada (talk) 18:04, 8 August 2024 (UTC)

Btw, I am not arguing Causation A which you accuse me of doing. I don't have sources for that, and never implied that immigration raised the overall rate of crime anywhere. Zilch-nada (talk) 18:05, 8 August 2024 (UTC)

I am not saying you did, I am saying that is how what you wrote could easily be interpreted, which is why the ordering is very important here to say from the outset that causation has not been proven before listing correlation studies and statistics (which I agree are probably worth including) Superb Owl (talk) 18:11, 8 August 2024 (UTC)

The "Worldwide" section implies that there is often a positive being immigrants and rates of crime "Even if higher crime is reported among immigrant communities, that could simply reflect a correlation..." The notion - and very often fact - of higher crime being reported among immigrants (regardless of their effect on overall crime, causation B) - do you seriously not think that (causation B as I described it above) - should be included at all in the lede? You have gutted almost every mention of B.

BTW, "simply reflect a correlation". Correlation in which way? Between immigration and overall crime (A), or immigrants and their respective crime rates (B)? Because the former - as you have already described - doesn't show much causation - but the latter, as the "Worldwide" section states, describes the causation (B) of immigrants committing higher crime rates; i.e., due to disproportionately locate in deprived areas where crime is higher (because they cannot afford to stay in more expensive areas) or because they tend to locate in areas where there is a large population of residents of the same ethnic background. So a causation is described that explores often higher rates of crime among immigrants. So why gut any reference to that? Zilch-nada (talk) 16:31, 10 August 2024 (UTC)

Thanks for pointing that out - I have tried to clean that up the 'Worldwide' section a bit to make it clearer and more organized. I agree it should have some mention in the lead. I think we add a paragraph summarizing correlations as well as 'perceptions' section. Am feeling more comfortable in moving forward after finally getting around to organizing 'Worldwide' section a bit better Superb Owl (talk) 19:40, 12 August 2024 (UTC)

I agree that there is no reason to only mention the US in the lede. I'm not sure that "the U.S. has the most reliable data" and we should not assume that the US is the only country that matters, or is a typical one. Alaexis_¿question? 21:13, 15 August 2024 (UTC)

The OECD study covers most of the ones with good data in this article (and some with less high-quality data). Bigger countries tend to have bigger sample sizes and better data, though not always. The U.S. is intended as an example to illustrate the OECD study findings but there is certainly room to swap out two of those examples with other countries' studies Superb Owl (talk) 23:35, 15 August 2024 (UTC)

My concern is that the US and EU experience seem to be quite different. Alaexis_¿question? 21:57, 16 August 2024 (UTC)

I am sure there are differences but do not see any evidence of that in reliable secondary sources so far Superb Owl (talk) 22:03, 16 August 2024 (UTC)

The big difference is that in the US the migrants generally commit fewer crimes than natives whereas elsewhere it's the other way around generally. See the Immigration and Crime: An International Perspective article that I've added which explicitly calls the US an exception.

At the same time it's true that studies haven't found a causal link between migration and crime even in the countries where migrants commit more crimes.

Scholarly sources like this one offer a more nuanced picture and we should use them rather than newspaper articles. Alaexis_¿question? 09:35, 20 August 2024 (UTC)

Little clarity on scope of this article

As I conversed above with @Superb Owl - who recently made massive changes in the article -, we seem to have different view of how the article should begin. Simply, the uncontroversial opening states,

• Immigration and crime explores whether there is a relationship between criminal activity and the phenomenon of immigration.

Now, that should quite clearly encompass a (A) potential relationship between immigration and overall crime rates - i.e., does a rise in immigration increase (causation) overall crime - or not. But I think it is also quite clear that "relationship between criminal activity and the phenomenon of immigration" also refers to, for instance, the rates of crime being greater - or lesser - in immigrant groups compared to other groups (B). The causation for this could be due to poverty, racism, police brutality, among other things, that, in some countries whereby immigrants commit higher rates of crime. So,

• A: Immigration and its effect on overall crime
• B: Rates of crime among immigrants, e.g., compared to other groups.

A and B are two clearly different yet necessary aspects of this article. But @Superb Owl got rid of references to the latter, describing it as correlation-and-not-causation. But that is a different scope; considering B does not consider the causation on overall crime, but the causations of immigrants comparatively committing higher-or-lower rates of crime. I.e., a different scope of causation. I don't see why the latter point should be excluded; i.e., "rates of crime among immigrants" and how it relates to "a relationship between criminal activity and the phenomenon of immigration". Zilch-nada (talk) 16:42, 10 August 2024 (UTC)

I would suggest we first continue to work through the body to continue cleaning-up individual sections before focusing on the lead. There are still issues of synthesis, original research, and selectively citing articles in a misleading way. Superb Owl (talk) 16:51, 10 August 2024 (UTC)

@Zilch-nada, both are in the scope of the article. Alaexis_¿question? 09:39, 20 August 2024 (UTC)

I think we disagree on what the meaning of the correlation in immigrants being overrepresented in crime statistics means. It previously was implied to mean that immigrants commit more crimes by being given undue weight (and improper synthesis). I removed that misleading language and poor sourcing.
We all agree that we need to discuss how some countries have crime statistics with proportionately more immigrants, but the reason most studies cannot establish causation that immigrants commit more crimes is many and also need explaining in that paragraph (see Worldwide section for the long list of reasons why correlation does not equal causation). We are much closer to have sufficiently robust sections in the article to start attempting to summarize in the lead and am happy to engage in that now given where we are Superb Owl (talk) 20:54, 20 August 2024 (UTC)

Europe

Currently the Economist's article is used to support the following statement in the lede

“ In Europe, there is also little connection between immigration and crime despite claims by some right-wing parties. ”

The Economist is a newspaper and not a peer-reviewed journal. It's generally a reliable source and normally it wouldn't be much of a problem. However in this case I believe we need a better source. The Economist article was published 6 years ago and a lot of research has been produced since then. Some of it is mentioned in the article, some of it isn't (just a couple of examples arriving to different conclusions [1] [2]). So I believe that we need to use a newer scholarly source for such sweeping statements, and if it's not available we should be more cautious. Alaexis_¿question? 21:38, 16 August 2024 (UTC)

The first is a primary source just of Germany (2015-2016) and the second a primary source just of refugees from Syria in Turkey - WAY too specific to make or dispute any generalizations. If you have reliable secondary sources that are more recent and contradict The Economist, that would be really helpful context. Superb Owl (talk) 21:52, 16 August 2024 (UTC)

These are secondary sources ("thought and reflection based on primary sources, generally at least one step removed from an event") and not primary ones ("original materials that are close to an event, and are often accounts written by people who are directly involved").

Again, a 6-yo newspaper article is not a great source for the current scholarly consensus. If we don't have good sources for that we should simply not say anything. Alaexis_¿question? 21:56, 16 August 2024 (UTC)

Just took it out of WP:Voice for now as you are right that we should be careful to make generalizations on one source. But it is still the best non-US source we have and so I think it belongs in the lead
And those are still primary sources with very limited scope of study - way too narrow to include in a lead or inform this conversation. Maybe worth including in the body, but even then a meta-analysis is what we really need Superb Owl (talk) 22:00, 16 August 2024 (UTC)

No, I'm absolutely not suggesting to include the sources I've mentioned in the lede (even though they are not primary). I doubt that a 2018 newspaper article is the best we've got, let me try to find better sources. Alaexis_¿question? 06:58, 20 August 2024 (UTC)

OECD result

The lede includes the following statement

“ A 2023 study of 30 OECD countries found no statistical link between immigration and crime. ”

This is technically true but there are other issues with it. The lead should summarise the article in a neutral way. Why do you think that this article published in a relatively low-impact Croatian journal (Economic Research-Ekonomska Istraživanja) represents the scholarly consensus? Alaexis_¿question? 21:53, 16 August 2024 (UTC)

It has 26 citations in the past year which is notable but that may be a better use of an 'additional sources needed' flag Superb Owl (talk) 21:58, 16 August 2024 (UTC)

Verification needed tags - Denmark

They were added en masse with no explanation in the reason parameter and no comments at the talk page. What are the reasons to doubt the sources there? Alaexis_¿question? 09:38, 20 August 2024 (UTC)

language, history of misuse of sources about Denmark in lead and elsewhere Superb Owl (talk) 19:19, 20 August 2024 (UTC)

Attribution

I'm not sure the attribution is warranted here. But if it is, then it's much more warranted for an article in Economic Research-Ekonomska Istraživanja (also an economic journal but much less prominent [3] [4]). Alaexis_¿question? 19:03, 20 August 2024 (UTC)

Fair enough - let's put back in WP:Voice. Just double-checking on econ journal relevance
I changed attribution to 'A 2024 article.' It does not argue to be a definitive meta-analysis of studies but is a single analysis that reaches rather weak conclusions on a limited set of countries, finding effects mostly in smaller european countries. It also has trouble isolating variables or putting into broader context its findings, admitting as much. This is a useful primary source, but not one we should include in the lead with only 2 citations so far, neither of which evaluate the paper or its claims. The lead should be reserved more for meta-analyses (if we can find them) or conclusions supported in the individual sections. Superb Owl (talk) 19:20, 20 August 2024 (UTC)

Yeah, that makes sense. The low citation count doesn't mean much for a recently published article. If you can find a better article we can surely add it.

I'm sorry for repeating it, but it's not a primary source in the sense this term is used on Wikipedia (WP:PRIMARY): Primary sources are original materials that are close to an event, and are often accounts written by people who are directly involved. Census is a primary source. A scholarly article that investigates a link between X and Y using various primary sources is a secondary source. Alaexis_¿question? 18:55, 21 August 2024 (UTC)

True it was published in 2024, but in 'winter 2024' meaning 8 months ago. The July 2022 article (25 months ago) from the less reputable journal has 26 citations already. Kinda hard to compare but hopefully we will get more anlaysis of their findings soon.
Disagree re: primary vs. secondary sources, per: "For example, a review article that analyzes research papers in a field is a secondary source for the research." Because they are conducting an original analysis of data (not analyzing much research), the claims about their findings of their analysis is primary in this context. A meta-analysis would be secondary. Superb Owl (talk) 19:03, 21 August 2024 (UTC)

Marie and Pinotti

The article's conclusion is very straightforward

“ On one side, in most countries—with the notable exception of the United States—immigrants exhibit a disproportionate involvement in criminal activity compared to natives, as measured by the relative incarceration rate of the two groups. In addition, certain kinds of immigrants, including young and less-educated men and those with irregular legal status, display a much higher propensity to commit crimes than those with documented status. These factors would seem to suggest a positive link between immigration and crime. On the other side, studies designed to measure the effect of immigration inflow effects on local crime rates do not, in general, find any detectable causal effect of immigration on local crime rates ”

We shouldn't replace the conclusion with cherrypicked facts from the article. Alaexis_¿question? 19:13, 21 August 2024 (UTC)

The conclusion is what seems cherry-picked after looking at the data - it makes it sound like the US is the only outlier (what about NZ, AUS, UK, and the dozens you cannot see because of the chart clutter). The article does not even attempt to say how many countries fit this correlation that they found. It also does not try to contextualize the numbers in any more statistically significant way such as the percentage of people who live in countries where immigration is associated with crime. Denmark and Sweden are technically more countries (2) than the US even though their population is far less. There are many statistical issues with the conclusion and I do not think it should be given very much weight in Wikipedia. Superb Owl (talk) 20:50, 22 August 2024 (UTC)

This is your opinion. If there are reliable sources that dispute this finding, we can add them to the article.

I've added a note that other Anglo-Saxon countries also behave differently.

The chart shows the prison population and not crime rates, so you can't say that the conclusion is wrong based on it. In any case this is WP:OR (analysis or synthesis of published material that reaches or implies a conclusion not stated by the sources). Alaexis_¿question? 20:55, 22 August 2024 (UTC)

The prison population is the variable referred to in this assertion that they used to assess crime rates. I added that to the text to be precise - they did not evaluate crime rates, just prison populations. I hear you on wanting it rebutted by secondary sources but without being able to see the data, I do not want to put much weight on this paper. And as mentioned before, this article has not been around for more than a year and has not had time to be rebutted, included or excluded in other articles so it's premature to put a lot of emphasis on its original analysis. Superb Owl (talk) 20:58, 22 August 2024 (UTC)

Well, I can nitpick pretty much any article in this field and find gaps there because it's not exact science and one can always find things the authors didn't consider.

I'll try to find more sources confirming this. Alaexis_¿question? 19:56, 23 August 2024 (UTC)

Charts

The charts listed in certain sections offer raw data that has not been peer-reviewed, no attempt has been made to isolate other causal variables or even contextualized for the variables that are not. These charts, in my opinion, provide undue weight to the narrative that immigrants create crime, which has yet to be proven. We should not have any charts that make that implication at all, but if we do, we should have more charts showing that crime has not increased due to immigration. Superb Owl (talk) 20:47, 22 August 2024 (UTC)

Danish charts

Per WP:PRIMARY, such sources can be used on Wikipedia provided we do not interpret them ourselves and do not use them excessively. This is clearly not the case here, there are plenty of secondary sources in the section on Denmark which discuss this data. Alaexis_¿question? 20:51, 22 August 2024 (UTC)

and per WP:DUE they cannot if they overemphasize one position, especially when it is a minority view. Superb Owl (talk) 21:42, 22 August 2024 (UTC)

The chart doesn't "overemphasize one position" and it's not a minority view. These are the numbers released by the Danish statistical service and there are no other numbers about the levels of crime by the country of origin in Denmark. Different scholars interpret these numbers differently and this is precisely what is discussed in the section.

Wikipedia is not WP:CENSORED. Alaexis_¿question? 19:52, 23 August 2024 (UTC)

These are raw data, not refined to isolate any variables or contextualize it. We can fill up a section with raw data but that is not the question that people want to know - they want to know whether immigration increase crime and if so by how much. Even if immigration does increase crime, there is a lot of evidence that the amount is not by as much as is implied by charts like that. That is why it is WP:UNDUE Superb Owl (talk) 19:56, 23 August 2024 (UTC)