Talk:List of countries by intentional homicide rate

Note: See table updating info: /LibreOffice Calc instructions and /Excel instructions.

map no longer correlates with data in the table

— Preceding unsigned comment added by 2603:8001:a702:13a8:351a:90bc:15e4:a903 (talk • contribs) 29 May 2022 (UTC)

Include multiple charts

There is no law restricting list articles to only the major topic, and all else is verboten.

• List of countries by firearm-related homicide rates#Charts and graphs
• List of U.S. states and territories by incarceration and correctional supervision rate#Comparison with other countries
• List of U.S. states and territories by incarceration and correctional supervision rate#U.S. states compared to countries
• List of U.S. states and territories by incarceration and correctional supervision rate#Female incarceration in U.S. states and countries
• United States drug overdose death rates and totals over time#Comparisons to other countries in Europe
• United States drug overdose death rates and totals over time#Rate timeline by race and ethnicity

--Timeshifter (talk) 15:33, 9 November 2024 (UTC)

Of course there's no "law", nor policy, but there are guidelines that play into this, for example stand alone lists and list selection criteria. Consistency in focus in a "list of" article should be self-evident, because the title is a pretty clear restriction and definition of what exactly it contains, unlike articles that delve into broader topics.

Your own previous example of List of countries by household final consumption expenditure per capita demonstrates this pretty clearly: it has a rigorously defined set of parameters that result in the data the list represents. Those final data are affected by an enormous number of other co-variables (resource availability, political structure, political corruption, societal corruption, population density, countless others) - but those variables aren't presented along with the list for good reason. A 'list of something' should confine itself to something for which a list is presented (yes, I'm intentionally using a tautology).

I would argue that in the selected group of "list of" articles you've presented, some of the graphs there are also inappropriate. For example, "List of countries by firearm-related homicide rates" (my emphasis) contains a circle chart of total firearm death rates combined - homicide, suicide, and other. Suicide and homicide are vastly different phenomenon, and the list contain no data on suicides. Honduras is the biggest circle in the chart - but Honduras has a very low overall suicide rate; South Korea isn't even represented as there are virtually no firearm homicides or suicides there, but it has an enormous overall suicide rate, by far the highest of any 'high income' country. It also has a chart of mass shootings, which are an extremely rare phenomenon (and subject to wildy different base definitions), and which don't easily correlate with firearm-related homicide rates - unless we carefully craft in even more parameters that aren't included in the list in order to present that chart.

The argument has been made in the past that "but people are interested in it" - which is nice but if that were a reasonable rationale, it would encourage a mountain of digressive (but "interesting") information added that is not presented in nor directly relevant to the list. An article entitled "Gun violence" covers the vast motives, cultural influences, confounders, public opinions, political approaches, etc. of a broad topic. It's obviously appropriate there to cover the many possible intersections of data that are relevant to a broad topic, including with graphs and charts. When the very title of an article implies that the content is restricted to a list of X, it's prudent to not digress outside of the data presented, as it's a data-centric article, not analysis-centric. 'See Also' exists for this very reason. cheers. anastrophe, an editor he is. 21:33, 9 November 2024 (UTC)

I don't see anything in your links that blocks or prohibits any of the notable things I linked to previously.

List of countries by minimum wage#OECD

That is a table using a subset of the country list. The OECD subset. There is no reason the homicide list can't include a highest median wealth subset. In table and/or in image form. --Timeshifter (talk) 00:06, 10 November 2024 (UTC)

Please read more carefully, I was quite clear in what I wrote that there isn't anything that blocks or prohibits them, but there are principles of consistency within what an article ostensibly is about. My argument is for content consistent with the data presented. Does the list this article contains have a column for where each country stands in wealth indices? It does not. It's a derivative, and is ultimately WP:SYNTHESIS that just cruds up a very straightforward and unencumbered List of countries by intentional homicide rates with sociological/political factors that aren't present. There are more appropriate articles discussing those topics already in existence, where those graphs would be perfectly appropriate. There are good faith, strong opinions concerning to what degree sociological and political variables affect criminality. They are detailed in articles that exist precisely for that information. I see a political agenda in adding easily manipulated data (via personal choice) that implies - whether intentionally or accidentally - causality or explanation. We should stay far, far away from appearing to endorse factors not present in the list.

As far as the OECD list, what you say can pretty easily characterize what's known as "mission creep". Why complicate this article which has ONE topic with other factors? Create a new article (assuming one doesn't exist) that lists the homicide rates against OECD nations. But no subsets. Absolutely not, as that's cherry-picking and falls right into the problem I'm addressing. Frankly, I've never even heard of anyone suggesting a correlation between minimum wage and homicide rates. cheers. anastrophe, an editor he is. 04:56, 10 November 2024 (UTC)

I think I've avoided the issue by uploading a new chart, affectionately labeled "Chart C", . It's directed to the 15 countries with the highest homicide rates, without depending on selection criteria other than (highest) homicide rate. I lean toward the inclusive side, that important measures such as income can be used as an inclusion criterion, and I think Chart A is more interesting and informative than Chart C, but I can live with either in this article. —RCraig09 (talk) 05:16, 10 November 2024 (UTC)

I would lean towards the inclusive side too, but the beauty of Wikipedia is that there's no real limit (other than certain policies) on the number or kinds of information that can be presented in individual articles. I'd have zero issue if there were an article List of countries by intentional homicide rates and GINI coefficients (or OECD or whatever strike's one's fancy), and it included the charts/graphs discussed here - though I'm always wary of the use of subsets of datasets. Though that can intrude on readability - e.g. the graphs immediately adjacent to List of U.S. states and territories by incarceration and correctional supervision rate#U.S. states compared to countries lead me to want to buy a 55" ultra-widescreen monitor flipped to 90 degrees. cheers. anastrophe, an editor he is. 05:37, 10 November 2024 (UTC)

Anastrophe. About List of countries by minimum wage#OECD. There is no WP:SYNTHESIS happening. It's just a common subset of countries used in many list articles. They are countries in the Organisation for Economic Co-operation and Development.

As to charts with homicide rates and totals for high income/wealth countries, that is just another subset. If you feel it slants the article, that is just your interpretation. People can interpret the data any way that they want.

Feel free to add additional subsets. I and others are curious about them all. That is how to meet WP:NPOV. More angles and "sides", not less. Censorship is not a solution to get to WP:NPOV. --Timeshifter (talk) 17:07, 10 November 2024 (UTC)

I'm not sure why I'm being misunderstood and I tire having to re-explain what should be easily understandable. If you add a graph to this article using minimum wage per OECD coupled to the homicide rate data, it absolutely is synthesis, because this article lists only countries and homicide rates, and addresses no other subjects. This article describes no point of view. Introducing metrics that do not exist in this article and conjoining them with the data that this article provides is inherently synthesis. It's not an extension of the data here. It's new data, outside the very direct scope of this article.

Your comment "As to charts with homicide rates and totals for high income/wealth countries, that is just another subset." is specious. It absolutely is not a subset of any data in this list. Again, please show me where this list includes data on high income/wealth countries. Is there a hidden column for it? (I'm being facetious). You can't call a metric that isn't in the dataset a "subset" of the data presented. As for the notion that there are "More angles and sides", no, there are not, actually. This article doesn't engage in any discussion of political, social, wealth, population density, race, shoe color, number of electric cars/100k, or any other coupling. This article is:

1. A list
2. of countries
3. by homicide rates
4. and homicide counts (which are the direct source of the rates)

Show me where in this article a 'Point of View' is described/discussed/presented via reliable sources. Hint: It describes no point of view, the only description is of how the data is defined, without interpretation. You may personally be interpreting the article and data through the lens of a particular point of view (which is rightfully the domain a reader controls) but that POV is absolutely not detailed in this article or the data it presents, because this is a non-POV article consisting of:

1. data
2. a description (not an interpretation) of the data
3. nothing more.

If you wish to create new list articles or other articles that include metrics beyond those in this article, by all means, do so (although almost all of the topics described in this discussion are already present in the many articles on crime, homicide, wealth or lack of same, the delta of all three, and more). Attempting to 'guide' readers to a point of view in an article that is otherwise strictly data with no interpretation or point of view (reliable sources, remember), is not NPOV, it is pushing a POV.

My comment about using subsets of OECD data in graphs/charts has to do with cherry-picking. There are only thirty-eight countries in the OECD. There's no material reason (size, readability, density) to limit the number of countries compared. The only reason to exclude any of the countries would be "guide" the reader, because the OECD itself has no formal subgroups within it. Any subsets are inherently interpreted.

The state of this article before the new graph was added was "pristine". One subject - homicide rates by country - and a list of those data. The other images in the article are not interpretive of the data - no interpretive content of any kind. There are literally dozens of other articles where interpretive graphs and charts are appropriate, and indeed those articles have such graphs and charts already - and can host more of them, for that matter, because those article delve into the topic from the many different directions that this article...does not. Adding a POV to a non-POV article is self-evidently POV pushing. cheers. anastrophe, an editor he is. 19:34, 10 November 2024 (UTC)

Anastrophe. You still don't understand what I said concerning an OECD subset. This is the third time I am explaining it. See:

List of countries by minimum wage#OECD

In that country list there is a subset table just for OECD countries.

We could put a subset of OECD homicide rates here in the homicide list article too if we wanted. The homicide list already can be sorted by region and subregion. Are you against that too?

There are lists on Wikipedia and elsewhere broken up by race, gender, income, and much more. See:

https://www.google.com/search?client=firefox-b-1-d&q=break+up+lists+by+race%2C+gender%2C+income

RCraig09 agrees with me: "I lean toward the inclusive side, that important measures such as income can be used as an inclusion criterion, and I think Chart A is more interesting and informative than Chart C, but I can live with either in this article."

I can live with both being in the article, and more. In that case some charts could be placed in a gallery at the end of the article. --Timeshifter (talk) 13:05, 11 November 2024 (UTC)

Yes, I see what you mean re minimum wage and the subset of OECD data. But in that case, the list is about wealth/income/earning data, and the OECD is the Organization of Economic Co-operation and Development, so the subset truly is a subset of the data presented. There is no column here of GINI/HDI etc. So while it's a subset here, it's a subset based on data that's not present in the list. You could certainly add an OECD homicide rates list, but the article title would need to be updated to show that the content is broader than just homicide rates. cheers. anastrophe, an editor he is. 21:31, 11 November 2024 (UTC)

A

B

C

D

The existence of the source for Chart D ("Violent Death Rates: The US Compared with Other High-income OECD Countries, 2010") proves that parsing homicide rates by national income (or "developed nation" status, or "major nation" status) is a rational and meaningful choice for summarizing a list, and is not editorializing on our part. I don't know of a WP policy or guideline requiring list articles to be "pristine". If anything, the text listing's failure to include an "income" category is the shortcoming, not the inclusion of a data-selective graphic. Of course we do have to be careful not to stray into truly irrelevant selection criteria (e.g., selecting countries that have exactly seven letters in their names). However, lead images necessarily cannot convey all the information in a list, so I think that charting a rational, objective selection of a data subset is permissible. Adding Chart C to the article makes the issue somewhat moot here, but I think that going forward, charts including reasonably chosen data subsets should be allowed, and even the norm. —RCraig09 (talk) 17:59, 11 November 2024 (UTC)

"If anything, the text listing's failure to include an "income" category is the shortcoming, not the inclusion of a data-selective graphic." By that rationale, it is a greater shortcoming that it lacks columns breaking down the homicide rates by race, age, and gender of both victims and perpetrators; type of weapon used; relationship or lack thereof of victim/perpetrator; and more. Wealth/income levels is only one type of(*) correlative parameter, and definitely not as "reliable" a correlation as some of the ones I just listed.

(*) I say 'type', because there are dozens of differing selectors to represent wealth/income. See List of countries by income inequality (with multiple different measures within) List of countries by Human Development Index, List of countries by planetary pressures–adjusted Human Development Index, and on and on. Who chooses which is the 'best' correlate?

This article has existed in this form of simply country/rate/count, for roughly eighteen years. It's a clear and concise set of data and it presents no interpretive data. I strongly object to adding new correlative parameters, when it would be far more appropriate to create a new "list of countries by intentional homicide rate and wealth indices" or some such.

This article is an eminently clear and concise presentation of data, with no POV correlations, as it is now, and as it has been for nearly two decades; I've referred to it countless times over that interval, having a longtime interest in crime/criminology. Make a new article if you want to guide reader's intepretation of the data via a specific correlation. There's no good reason not to, and that way readers can still get the homicide rate information here, unencumbered.

If this is a fait accompli, then I guess after you guys have added what columns you want to the list (and re-titled it, which I believe would be necessary) I'll create a new "List of countries by intentional homicide rate". Kind of a goofy path, to have to recreate an article that already exists, instead of creating a separate (and accurately titled) article with whatever desired correlative data other editors want to add.

But until then, leave out the correlative graph, as it's derived from data that's not an included parameter in the list, and therefore isn't appropriate in its current incarnation; it pushes the notion that wealth is the most meaningful/accurate correlate for homicide rates, which it is not. cheers. anastrophe, an editor he is. 21:22, 11 November 2024 (UTC)

@Anastrophe: Actually, some of the things you list would make for meaningful comparisons, and an array of charts showing such factors would be appropriate in this top-level (least specific) article. If not wealth (or similar measure), what do you think is "the most meaningful/accurate correlate for homicide rates"? —RCraig09 (talk) 21:47, 11 November 2024 (UTC)

That falls right into the problem of editors making decisions of which correlates are "important", when correlations inherently can't be nailed down, and are easily abused to insert a POV not supported by the strict, limited data presented. My personal opinion is that age and gender of victim/perpetrator are far and away the more meaningful/likely-predictive correlates. But this isn't a "top-level (least specific)" article - it is very specific in what it presents, unlike homicide, familicide, suicide, which are top-level and under which the many more-specific articles fall (though there's no distinct heirarchy, obviously).

I'll take a snapshot of the wikitext of the article and prepare to make a new article when the time comes. cheers. anastrophe, an editor he is. 22:26, 11 November 2024 (UTC)

I have been editing this article since Oct 18, 2008. The reason more breakdowns haven't been added is because until recently it has been difficult to pull them quickly from the dataset. There are many breakdowns listed in the dataset:

• https://dataunodc.un.org/sites/dataunodc.un.org/files/data_cts_intentional_homicide.xlsx

I only recently figured out a faster method anybody can use without paying for spreadsheet software. Those other breakdowns can be pulled out by adapting these instructions:

• Talk:List of countries by intentional homicide rate/LibreOffice Calc instructions

Some of the breakdowns are listed at the top here:

• https://dataunodc.un.org/dp-intentional-homicide-victims

Country list articles don't change their titles every time a new breakdown is inserted in the article. That would be silly.

These GDP list articles have 4 different sources (IMF, World Bank, CIA, and UN):

• List of countries by GDP (PPP) per capita
• List of countries by GDP (nominal) per capita

We don't put the sources in the title in this case. That would be silly, because sources may change or disappear. As apparently has happened here:

• List of countries by wealth per adult.

UBS has apparently stopped putting out this info. At least in an accessible way that Wikipedia editors can use. See its talk page. Before UBS was the source Credit Suisse was a source. --Timeshifter (talk) 16:44, 13 November 2024 (UTC)

Homicide rates are definitely, though maybe not dramatically, affected by income.

User:Timeshifter, I'm not seeing a "bottom line" with your 16:44 post. How does it affect this list article?

To all: look at the new chart to right. If someone went to the trouble of adding GNP-per-person to the textual table, would this chart merit inclusion? I won't think further about spending time adding a column of text to a huge chart, unless there's an indication of approval for the chart. —RCraig09 (talk) 03:49, 15 November 2024 (UTC)

No matter how hard I squint, that chart is barely more than a shotgun pattern, no pun intended. You have very low GNI-per-capita countries with very low homicide rates, and very low GNI-per-capita countries with very high homicide rates. Along the 1.0/100k rate line, there's a string of maybe 24 countries ranging from about $2.5k GNI-per-capita to just under $100k GNI-per-capita. "Definitely" is a misnomer when applied to correlations, weak or otherwise.

I reiterate my previous objections. I reiterate that there are many, many other articles, already in existence, list and otherwise, where this would be appropriate and meaningful. Why must we make attempts to "guide" the reader towards conclusions drawn from information not discussed at all in this article? 'See Also' exists for a reason. cheers. anastrophe, an editor he is. 04:45, 15 November 2024 (UTC)

The trend "line" generated by Microsoft Excel—presumably least-mean-squares (LMS) approximation—removes the need for "squinting". It goes from one order of magnitude homicide per two orders of magnitude income: a definite mathematical correlation though with substantial statistical variance.

? Your "24 countries" statement is perplexing, given that ~80% of the 166 countries are in that range and therefore largely determine the overall LMS trend; data points don't have to be "near" a trendline (if that's what you meant) to determine correlation of underlying data. I'll wait for input from other editors. —RCraig09 (talk) 06:30, 15 November 2024 (UTC)

My previous point was concerning whether country list article titles needed to be changed depending on changes in breakdowns and sources. They don't need to be changed when breakdowns and sources change. It is a lot easier to leave the title alone.

I think a scatterplot based on median wealth or median income would have a much stronger correlation. Average Gross national income (GNI) per capita is kind of meaningless in many cases in countries with high income inequality.

It is appropriate and meaningful to add these charts to list articles. Readers appreciate them. Readers see the data without having to wade through long articles to find it. They can interpret the data how they want. If readers want long prose "explanations" of the data from the right, left, and center they can go to longer articles. Personally, I prefer my data without the spin. --Timeshifter (talk) 18:54, 15 November 2024 (UTC)

Uploaded 25 Nov 2024. Median disposable household income (measured after taxes and benefits).

@Timeshifter: I took up your 18:54 suggestion and created the chart at right focusing on median disposable household income. I think you were correct that there is a closer correlation (less variance about the trendline) than in the earlier chart (GNI per person). The trendline slopes from the two charts look nearly identical but that is not particularly meaningful because the two horizontal axes (two income types) are distinct. —RCraig09 (talk) 22:18, 25 November 2024 (UTC)

For the GNIpp data, R² = 0.1693. For the MedianDHincome data, R² = 0.3614. In general, R² is the Coefficient of determination which is "the proportion of the variation in the dependent variable (homicides) that is predictable from the independent variable (income)". So here, homicide rate does indeed track median-income data more closely, as you predicted. —RCraig09 (talk) 23:44, 25 November 2024 (UTC)

@RCraig09: Wow. Great work! I see this from the file page: "49 nations were included, based on countries that were in both sources above." That's good to know. That means no cherry picking of sources was done, and makes the correlation more interesting. I suggest putting a 2-column table on the file page. With the 49 countries, and columns for income and homicide rates.

Also, can you put the finished spreadsheet in a shared folder of finished spreadsheets on your Google Drive page. I want to open it in LibreOffice Calc.

--Timeshifter (talk) 08:36, 26 November 2024 (UTC)

@Timeshifter: This link should work for a one-file folder I just created. I think you must have a Google Drive account to access. I couldn't remember a way to explicitly guarantee public access so let me know if you encounter a problem accessing. —16:43 . . . Be aware that I used a text editor to make various additions to the .svg file after pasting the basic svg code from the spreadsheet, so what you will get from the spreadsheet will be different. —RCraig09 (talk) 16:47, 26 November 2024 (UTC)

They look very different to me. But maybe they are supposed to look that different. No way I can know without having a spreadsheet that produces an SVG file that has been uploaded without alterations. It doesn't have to be the current version of a file.

I looked at the spreadsheet in Google Docs too. The SVG file from it looks the same as the one from LibreOffice Calc. So that's probably a sign that both Calc and Google Docs are working correctly with your spreadsheets? --Timeshifter (talk) 04:30, 27 November 2024 (UTC)

Yes, they are supposed to look different: I made additions (background colors, +gridlines, logarithmic numbers) before the first upload to Wikimedia. I've just uploaded a PNG screenshot called "20241126 Version direct from spreadsheet (before manual changes)" to this link. It should hopefully be the same as what Calc yields. —RCraig09 (talk) 04:57, 27 November 2024 (UTC)

I have Google Drive for an alt name I use.

The SVG code produced from LibreOffice Calc and Google Docs is exactly the same:

https://en.wikipedia.org/w/index.php?title=User%3ATimeshifter%2FSandbox278&diff=1259822365&oldid=1259821369

The PNG file is the same as the SVG file, except for sizing. I guess you had to pick a size.

I want to nail this down. Can you upload to Google Drive the SVG file produced from Excel for the unaltered spreadsheet. I can then upload the SVG code to my sandbox:

User:Timeshifter/Sandbox278

Then I can compare the text versions. If there is no change, then we know that people can also use LibreOffice Calc and Google Docs. --Timeshifter (talk) 07:23, 27 November 2024 (UTC)

The 3.7 Kbyte file "20241125 Homicide rate vs median disposable family income, by country - last version from spreadsheet.svg" is at this link. When you first click on it, Google Drive shows the "rendered" image, but when you download, you get the .svg text file itself. —RCraig09 (talk) 17:42, 27 November 2024 (UTC)

Thanks! The .svg text file from Excel is exactly the same as the .svg text file from Google Docs and LibreOffice Calc. See text file comparison:

• https://en.wikipedia.org/w/index.php?title=User%3ATimeshifter%2FSandbox278&diff=1259962223&oldid=1259822365

--Timeshifter (talk) 01:38, 28 November 2024 (UTC)

Microstates

Writing 0 for microstates is MISLEADING. If 1 murder takes place in the Vatican, it will immediately increase the rate to 125 per 100,000. And there were murders in the Vatican by the way. My proposition is to put "less than {and the rate for 1 murder in this country}". Or figuring out an average rate for, let's say, a decade. --95.24.74.171 (talk) 05:42, 1 January 2025 (UTC)

From a google search: "The last murder on Vatican grounds was on May 4, 1998, when Swiss Guard Vice Corporal Cédric Tornay shot and killed Alois Estermann and his wife, Gladys Meza Romero. Tornay then killed himself by shooting himself in the head"

See also: Crime in Vatican City.

Do you have a specific example of what you are talking about? --Timeshifter (talk) 10:18, 1 January 2025 (UTC)

@Timeshifter: OK, that means the murder rate in the Vatican is 12 per 100,000 population per year. --95.24.79.231 (talk) 02:44, 5 January 2025 (UTC)

Population:

https://www.worldometers.info/world-population/holy-see-population

In 2023 there were zero murders according to UNODC. Zero divided by population of around 500 equals zero. That's zero per 100,000 in 2023. That's zero per 500 in 2023. Zero, zero, zero. --Timeshifter (talk) 11:29, 5 January 2025 (UTC)

Pakistan's Number

It seems like the Pakistan number is off. Most likely a , for a . 141.89.234.16 (talk) 11:41, 20 January 2025 (UTC)

Thanks. I checked the source. I fixed it. You can too. Anybody can edit Wikipedia. --Timeshifter (talk) 00:19, 21 January 2025 (UTC)

El Salvador

El Salvador data has not been updated; the homicide rate for 2024 is 1.9 165.165.122.22 (talk) 19:08, 7 February 2025 (UTC)

Seems UNODC has a long time delay. Can add more recent data together with reliable source. HudecEmil (talk) 10:08, 8 February 2025 (UTC)

We only use UNODC data. That was decided long ago. See talk archives. We let UNODC do the vetting of the data. --Timeshifter (talk) 19:40, 8 February 2025 (UTC)