RSS
My discovery a couple of years ago of how Deaths of Exuberance (murders and car crashes) respond to headline events like Ferguson and Floyd would seem like the most important in the social sciences since Case and Deaton’s discovery of Deaths of Despair in 2015.
On the other hand, it remains one of the better kept secrets of the decade, probably for multiple reasons. Mentioning it in the media makes America’s sacred cows look less sacred, it makes The Establishment look less competent, and the discover is a Crime Thinker and if he’s right about such an important question, what else might he be right about?
The best I can come up with is to just keep working on displaying the data ever more self-evidently.
What are the best ways to show graphically how post-George Floyd depolicing contributed to the ongoing surge in homicides and traffic fatalities?
This graph shows the cumulative number of blacks who died by homicide for each of the 52 weeks of the year. Black homicide deaths were lowest in 2018, the blue line at the bottom, then slightly higher in 2019 (green line).
In 2020 (the crucial red line), homicides were running slightly higher again in the first 11 weeks of the year before covid became a big deal. The homicide rate worsened a little over the next ten weeks of lockdowns, then kinked sharply upward after George Floyd’s death.
2021 (black line) was very similar to 2020 after George Floyd. In other words, 2021 was just a continuation of 2020’s high plateau of black murders with the exception of now period of lower murders in 2021 before May 25 as there was in 2020.
2022 (gray line) started off as bad as 2021, but eased off over the summer. But the cumulative figure was still somewhat worse than in 2020 when the CDC’s death data runs on at the end of September 2022. (The CDC imposes a 6 month lag on death from “external causes” data in case investigators change their mind from say, car crash to suicide to homicide).
The black motor vehicle accident death figures are perhaps even more obvious:
The 2018 and 2019 curves are hard to distinguish because they are on right on top of each other.
The red 2020 line was running just above the previous years during the winter and during the covid lockdowns. Since less miles were being driven in March through May, that suggests bad driving among those out on the roads. But the big kind in the graph is in the weeks after George Floyd’s demise, during the nation-wide Mostly Peaceful Protests when the cops retreated and stopped pulling over bad drivers so much for fear of becoming the next Face of White Supremacy.
2021 (black line) continued the pace of the last seven months of 2020.
The gray 2022 started out as bad as 2021 but then eased off considerably in Q2 and Q3, and by the end of September was down to the 2020 level of deaths, but still far above 2018-19.
I also can plot deaths by month going back to 1999 (but only by week since the beginning of 2018):
This has the advantage of putting the 2020s in a much longer historical perspective. Wow, 2020 is crazy compared to the usual changes from year to year. A downside is that nobody can discern without a microscope can discern the difference between covid and George Floyd
And I can do the last 5 years by week, as I showed you yesterday:
English
Arabic
Chinese (Simplified)
Dutch
French
German
Hebrew
Hindi
Italian
Japanese
Korean
Persian
Polish
Portuguese
Russian
Spanish
Swedish
Turkish
Vietnamese
I don’t know. When does Tucker start here?
We had one guy--just one--speaking a bit of plain common sense "truth to power" on a mainstream media platform reaching millions of normies.
The minoritarian establishment badly wanted him gone ... and now he's gone.
The graphs you have been using so far have been just fine.
I particularly like the one that has date bars that show the beginning of Covid in early March 2020 and then Floyd’s OD with a second bar.
But here's the problem. The entire BLM "cops are killing blacks!" fraud is debunked by the single statistic that blacks are about 30% of police killings, but account (at the time) for 55% of homicides. (Homicide being a pretty good metric for the sort of situation "guy waving a gun around" call that gets cops showing up with weapons drawn and generates police shootings.) So police if anything are "under shooting" blacks relative to miscreants from other races. But that very clear stat, had zero traction.This simply is not about numbers--math, logic, reason. This is about this now dominant verbalist overclass religious ideology--minoritarianism--virtuous minorities oppressed by big bad racist white gentiles. With its sacred stories (the holocaust, etc.) of which the black experience in America--slavery, Jim Crow--is critical and foremost in the American rite.You aren't going to "debunk" any part of the narrative, with some time series behavior of blacks getting extra "vibrant". That blacks are oppressed by big bad whitey is a fact--a cornerstone of the massive minoritarian edifice, that justifies so much parasitism and looting and destruction.Replies: @Pixo
Individual city murder graphs with date bars showing would be good too. E.g., St Louis showing before and after Brown was killed.
This seems like a question for Andrew Gelman. He regularly goes on about how data visualization should be improved, and how just graphing all the data is better than summarizing it with a single statistic.
… what happened to blacks in 2001?
If you don’t want to see the data, no improvement in graphics will help.
Two thoughts come to mind. One, the key for the two events should be a line rather than a box, in order to co-ordinate with the vertical lines of the chart (events not trends).
Or, somewhat contrariwise, since both events really had durations, maybe they could be represented with diminishing shades, like so:
Your last chart is clearest, so perhaps this effect can be transposed there.
Past my bedtime now though 😉
Also, when considered all by itself, the idea that de-policing results in more black homicides is not particularly surprising or novel -- that's basically the already much-discussed "Ferguson Effect." So maybe you should market your meme as the traffic fatality corollary to the Ferguson Effect. But IMHO, you still need some causation evidence to make the whole thesis stick together. For example, one question I would ask is why are these deaths are still continuing to rise. If they are being cased by a mass black psychological response to the Saint Floyd hysteria, then shouldn't the effects be dampening over time as the Summer of Floyd recedes into the past? How do we know Floyd-ism has anything to do with it all, rather than some other complicated combination of social factors playing out over time, or hand guns, gas prices or sunspots?For example, does the increase in black homicides inversely correlate with the number of incarcerations of blacks for homicide? And does the increase in traffic fatalities likewise inversely correlate with traffic tickets written to blacks? A graph that shows such common correlation of homicides/traffic deaths to concrete metrics of de-policing metrics might be compelling to prove they are both symptoms of de-policing. With a little more causation evidence you might have something more viral -- as in, "Hey look de-policing actually has consequences (duh!), and this can be demonstrated by showing how it correlates with two different sets of bad outcomes in two very different domains."Replies: @anon, @Mike Conrad, @epebble, @Pixo
Same thing as happened to whites and Latinos: 9/11.
This isn’t as dramatic as one over time showing pre-Floyd rates would be, and it suffers distortions like putting Seattle on top.
https://www.statista.com/chart/23905/change-in-homicides-in-us-cities/
Visually, the cumulative graphs do not work nearly as well as the by-month graphs. Avoid using the cumulative metric since they do give a intuitively clear visual display of the information.
Years are a perfectly reasonable interval for bunching data, but year cumulative for displaying subyear--daily/weekly/monthly--data is basically artificial and pointless, and actually makes dramatic changes harder to see.
I can clearly see the inflection point at the end of May 2020. But then I've done calculus and can say "hey there's the inflection point". But sussing that out for most folks is harder.
Just show the weekly data--in in the last "blacks gone wild" chart and "wham!" the discontinuity is obvious to anyone--anyone who wants to see it.
The idea should be to make the effect as obvious as possible. To do this, plot per-month or per-week data rather than a cumulative plot. Also, plot ‘excess deaths’ rather than absolute deaths. That is to say, subtract off an average for that week or month over some number of years before Floyd.
My preference would be per-week data starting at the beginning of 2020 and finishing at the end of 2021, possibly with various newspaper/CNN headlines in boxes linked to points on the x-axis.
It would be interesting to try to pinpoint when the excess took-off. Was it actually the week Floyd died?
I can’t deny the evidence, but I find it difficult to understand the mechanism. What is going through the perpetrators’ minds? Presumably they believe that Chauvin killed Floyd, and that it was deliberate, and that it was because he was black, and that a significant proportion of police would like to kill black men too, but I don’t see how you get from that to shooting into parties full of black people.
I don't know if that was covered in How to Lie With Statistics, but it's a common criticism of op-ed graphs.
Similar tricks are discussed in How to Lie With Maps. Which kind of gave me the idea that we should adopt the pre-woke PC Gall-Peters Projection. That makes the Third World look much bigger than in Mercator projection, i.e., more of a threat. I can't believe its champions have never noticed this-- or maybe they have. You don't see Gall-Peters as much anymore.
You might even distort it even further, like putting Europe and North America on the edges of a more circular, fish-eye projection. "We have no room for all these people! We are full!"
https://originalmap.us/wp-content/uploads/2019/08/Word-Map_Gall-Peters_Original-Map-600x390.jpg
In the first two graphs, the legends on the right should be removed and replaced with text labels above each line in the graph itself (the colors of each line and label should match). This would greatly enhance readability. Covid onset and George Floyd’s death should similarly be indicated with text labels next to the vertical lines. I don’t know what program you are using to generate graphs, but the labels may be easiest to add afterwards with a graphics editor. The last graph would be better with text labels as well.
You didn’t discover this Steve, it was discovered by a black lesbian single mother who tweeted ‘Bix nood mufugga’ which contains all the data you subsequently plagiarised.
The vertical lines in the first two graphs should be removed because they apply only to 2020. Instead, add arrows and text labels that point to the correct dates in the 2020 line.
Yes. As Steve said, the chief problem Steve’s insight faces is that his target market does not want to believe it, for what are essentially religious reasons (though they wouldn’t say it that way).
Given that, the data-message has to be extremely direct and visually impactful, since any amount cognitive processing required to understand the graph is an excuse for the target market mentally to check out: “I don’t have time to interpret all of this”, “The effect looks small, whatever it is”, “Too much confounding information”, etc. They will seize (perhaps only unconsciously but therefore more powerfully) any excuse to avoid challenging their catechism. So no excuses must be offered.
Therefore the “cumulative” graphs are basically useless for marketing this bit of unwelcome news. The best that can be said for them is that the red line’s sudden departure from its old trend might remind older readers of that Pink Floyd album cover, but I don’t know how many of those people remain. The “cumulative” graphs have both the problem that they require considerable cognitive processing, and that for those few who are willing to do the processing, the effect—a change in slope from about 30° to about 35°—does not look too bad. Thousands of deaths are encompassed in that change, but visually it just does not look like much is happening.
By far the best graphs for marketing are the simple weekly time series where deaths are chugging along at a self-evidently “normal” rate and then a sudden discontinuity occurs which permanently changes everything. It’s like an EKG: [insert EKG sound effect] everything rolling along nicely, then suddenly, WHAM! Something serious happened. What was it? Zoom in on the date and there is Fentanyl Floyd and the Racial Reckoning. The very little cognitive processing that is required is done naturally since the chart tells an obvious story and the brain naturally wants to find out what this story is. The sudden change from one week to the next rules out other comfort-food explanations (“covid”, “Trump”, “racism”, “white supremacy”, Russians”).
While the long timeline covering 9/11 and the Ferguson Effect is interesting to cognoscenti here, it distracts from the primary message and offers adverse readers (everyone who matters) an exit before reaching the important conclusion. So cut that stuff out, at least initially. There has to be one story only and it has to be obvious. If they’re willing to see that, then later they can be let in on Ferguson Effect: “See, there was a sample run back in 2014, then they went whole hog in 2020”.
The bottom graph (the one from yesterday) is probably the most effective. Something very obviously went wrong in the 21st week of 2020. The x-axis is pretty busy though. Maybe just have three dates: the beginning, Floyd’s death, then the end. And put the dates in normal date format. Figuring out what “2020/22” means requires too much additional cognitive processing. Having the two vertical bars is good. It makes the false hypothesis (Covid) and the true hypothesis (Floyd) self-evident, but in the legend, instead of a matching fine vertical line, there is a big colored box, which once again requires adverse readers to make a cognitive leap to connect those colors with the very different shapes they match to. Either take “Covid” and “Floyd” out of the legend and put them directly on the chart, or at least make the little symbols in the legend match the fine vertical line shape they have in the main chart. (I know you’re probably using Excel, which has rather primitive graphing functions, but there are ways to English these things.)
For the second-to-last chart (Motor Vehicle Death Rates), consider starting the x-axis after 2010, and consider starting the y-axis at 5 instead of zero (I hope Achmed doesn’t read this comment!). I know Achmed considers the latter misleading, but the Prestige Press do it all the time and so it is what people are used to seeing. As the chart is now, the important part is off by the upper right edge. It needs to be front and center.
Oh, one more thing: the background grids in the third and fourth charts give a fun EKG vibe, but depending how the charts are reproduced, the grids can end up becoming darkened and noisy. Maybe pre-emptively lighten them somewhat or even remove them. The data must stand out strongly from the background clutter. (See Edward Tufte for a deep dive on this.)
I like the bottom graphs. I didn’t understand the new graphs at first glance. For a graph to be effective it needs to be immediately understandable.
You should extrapolate the pre-kink line so we can see the difference more clearly.
Definitely use the instantaneous week-on-week graph, not the cumulative graph.
Steve you should strike a devil’s bargain with a real academic economist, who could actually publish results like these in an academic econ journal. He goes ahead and publishes this stuff so that it gets out there from someone else and academia and the media can’t just 100% pretend it doesn’t exist (they will only be 99% able to pretend it doesn’t exist, like with the Deaton/Case results). He gets whatever credit comes his way, and you get the results out to a larger audience.
When that person’s colleagues, students, and Twitter followers start saying “Wait a minute, you got these ideas from that crime thinker Steve Sailer!” he can say “Who is Steve Sailer? I never heard of that guy, but YOU must be reading Steve Sailer if you think he has done something similar to what I’ve come up with.”
I particularly like the one that has date bars that show the beginning of Covid in early March 2020 and then Floyd’s OD with a second bar.Replies: @AnotherDad, @Bill Jones
Agree. Basically the last three graphs tell the overall story more than adequately.
The last graph in particular debunks the “Covid messed it up” excuse making and points the finger: In the wake of Saint George’s OD, the establishment amped up the minoritarian lying telling blacks “Whitey is oppressing you. And you’re right you should be really mad!” and sure ‘nuf blacks responding as you would expect.
But here’s the problem.
The entire BLM “cops are killing blacks!” fraud is debunked by the single statistic that blacks are about 30% of police killings, but account (at the time) for 55% of homicides. (Homicide being a pretty good metric for the sort of situation “guy waving a gun around” call that gets cops showing up with weapons drawn and generates police shootings.) So police if anything are “under shooting” blacks relative to miscreants from other races. But that very clear stat, had zero traction.
This simply is not about numbers–math, logic, reason. This is about this now dominant verbalist overclass religious ideology–minoritarianism–virtuous minorities oppressed by big bad racist white gentiles. With its sacred stories (the holocaust, etc.) of which the black experience in America–slavery, Jim Crow–is critical and foremost in the American rite.
You aren’t going to “debunk” any part of the narrative, with some time series behavior of blacks getting extra “vibrant”. That blacks are oppressed by big bad whitey is a fact–a cornerstone of the massive minoritarian edifice, that justifies so much parasitism and looting and destruction.
Nathan Cofnas said it well in a substack I linked to a few weeks ago:
“But wokism is protected by a taboo on recognizing the genetic basis of race differences. From early childhood onwards, we’re taught that taking hereditarianism about group differences seriously is a heinous moral crime. All psychologically normal people have been affected by this experience. Until recently, there was a similar taboo on questioning the Biblical account of creation. Darwin himself, when sharing his view that species are not immutable, wrote that it felt like he was “confessing a murder.” In 2023, taboo violators face not just internal psychological resistance, but a very real threat of punishment.
Again, people have to make a decision: we can either live with wokism or do what is necessary to stop it. Whoever wills the demise of wokism wills knowledge of race differences.”
-
It is easier to diagnose the problem than solve it. My best solution is to break the taboo as non-threateningly as possible, but without apology. Steve has done an exceptionally good job of this for decades and is a valuable model.
And the part you won’t like: picking your battles wisely, to the point of complete surrender on the JQ, and switching over to effusive support for Israel and adopting Christian Zionist language like “Judeo-Christian American values.” No better friend, no worse enemy.Replies: @Mike Conrad, @JimDandy
Agree.
Years are a perfectly reasonable interval for bunching data, but year cumulative for displaying subyear–daily/weekly/monthly–data is basically artificial and pointless, and actually makes dramatic changes harder to see.
I can clearly see the inflection point at the end of May 2020. But then I’ve done calculus and can say “hey there’s the inflection point”. But sussing that out for most folks is harder.
Just show the weekly data–in in the last “blacks gone wild” chart and “wham!” the discontinuity is obvious to anyone–anyone who wants to see it.
Losing Tucker from Fox is a real blow to us.
We had one guy–just one–speaking a bit of plain common sense “truth to power” on a mainstream media platform reaching millions of normies.
The minoritarian establishment badly wanted him gone … and now he’s gone.
I think a good approach is to lead with your final graph and emphasize in text that the timing indicates the more important cause (i.e. pretty much what you have been doing). One side point is that I think it might be more accurate to view Covid as preparing the fire (e.g. gathering kindling and wood and placing them) with George Floyd being throwing on gasoline and a match.
After that use the two prior graphs while highlighting in text that they place the shorter term trend in a broader temporal context and show how much the trends differ by race.
Another possibility is to focus on 2018-2022 and show year over year changes (I think percentages would be better than absolute numbers) in new versions of all three graphs. I think that would even better highlight the post-George Floyd step function and also bring out more the increased seasonality in black deaths which I am not seeing for the other races (except white motor vehicle deaths which have significant seasonality over time but relatively small post-Floyd increase in that). I am curious if that trend will persist.
P.S. Mike Conrad’s shading idea seems like it is worth a try. Then compare how it looks to the original versions and choose.
Thanks.
My preference would be per-week data starting at the beginning of 2020 and finishing at the end of 2021, possibly with various newspaper/CNN headlines in boxes linked to points on the x-axis.
It would be interesting to try to pinpoint when the excess took-off. Was it actually the week Floyd died?
I can't deny the evidence, but I find it difficult to understand the mechanism. What is going through the perpetrators' minds? Presumably they believe that Chauvin killed Floyd, and that it was deliberate, and that it was because he was black, and that a significant proportion of police would like to kill black men too, but I don't see how you get from that to shooting into parties full of black people.Replies: @Steve Sailer, @Reg Cæsar
Interesting idea for a graph of excess deaths. One problem is that calculating the excess is subjective for homicides because 2019 was higher than 2018 and Jan-Feb 2020 started out higher than 2019. It’s not indisputable what should be the projection for 2020.
On the other hand, for traffic fatalities, 2018 and 2019 were nearly identical so the average of them would make a fine projection for 2020 to use in calculating excess.
For example, the all-time murderiest day in Chicago history was Sunday, May 31, 2020, six days after George Floyd’s death, with 18 dead. I concocted a fictionalized account of what might have been going through somebody’s head here:
https://www.takimag.com/article/surviving-the-happiness-explosion/
Yes, a correlation. But that does not mean causation, as has been correctly noted before.
You say you loves data. Put your pattern recognition skills to the test to bolster your case.
https://elephrame.com/textbook/BL
Mine the data. Clearly define terms and criteria. Establish how many BLM protests were violent and murderous, along with how and why each protest met that criteria.
What you need to start tracking are Deaths from Depravity.
If there is a better word to describe what was done to this kid than evil I don’t know it.
https://thepostmillennial.com/trans-teen-died-from-vaginoplasty-complications-during-landmark-dutch-study-used-to-justify-child-sex-changes?utm_campaign=64483
I think that you’re mistaking secret with no one caring, including blacks. Tree falling in the woods kind of thing
Therefore, the best graph is, in fact, no graph at all.
Interesting… Ask Ohpra.
No, that doesn’t make any sense. Look at that spike. Was everyone in the Manhattan financial sector, onboard FakeFlight 93, and working in the Pentagon’s Office of Misplacing Billions of Dollars, black?
The black spike move is small from 27ish to 30ish. 3 extra deaths per 100K blacks.
White the Hispanic spike jumps from 9ish to around 17ish. 8 extra deaths per 100K Hispanics.
And the white spike is a jump from around 3 to 16ish. That's 13 points--compared to 3--and the white population is almost fives times bigger than blacks. So--ballpark--13/3 * 5 roughly 8 times as many whites killed on 911 as blacks. (More or less what you'd expect.)
This might be a little on the side of too cynical, but maybe just a bar graph with two bars in it showing the weekly average before and after May 25, 2020.
A less cynical version of this same idea would be side-by-side box and whisker plots of weekly averages before and after.
The thing is, the more opponents engage with the data, the better. If they have to come up with copes about the simplified presentation, then they can be shown followup graphs.
These gently sloping lines remind me of brown charts in Paul Samuelson’s “Economics”. Price goes down, demand goes up and Happy Equilibrium is achieved. Yay !!!
Vibrancy (and different brands thereof) is mapped in real time right there:
https://spotcrime.com/
What is strange – at the richest zip codes in the US: 94027 & 90210 vibrancy level is ZERO.
Those who bother to read “Data Science for Dummies” might consider digging up all those historical data and write an app with animations – “Evolution of Vibrancy in the XXI Century”
As tragic as these “deaths of exuberance” are, they fade into insignificance if plotted on the same chart as deaths by substance use which are also “deaths of exuberance” of another kind.
And they say Chicago pays no attention to Minneapolis.
My preference would be per-week data starting at the beginning of 2020 and finishing at the end of 2021, possibly with various newspaper/CNN headlines in boxes linked to points on the x-axis.
It would be interesting to try to pinpoint when the excess took-off. Was it actually the week Floyd died?
I can't deny the evidence, but I find it difficult to understand the mechanism. What is going through the perpetrators' minds? Presumably they believe that Chauvin killed Floyd, and that it was deliberate, and that it was because he was black, and that a significant proportion of police would like to kill black men too, but I don't see how you get from that to shooting into parties full of black people.Replies: @Steve Sailer, @Reg Cæsar
This can be carried too far. As restricting the graph to the top decile or less exaggerates incremental changes, making them look more dramatic.
I don’t know if that was covered in How to Lie With Statistics, but it’s a common criticism of op-ed graphs.
Similar tricks are discussed in How to Lie With Maps. Which kind of gave me the idea that we should adopt the pre-woke PC Gall-Peters Projection. That makes the Third World look much bigger than in Mercator projection, i.e., more of a threat. I can’t believe its champions have never noticed this– or maybe they have. You don’t see Gall-Peters as much anymore.
You might even distort it even further, like putting Europe and North America on the edges of a more circular, fish-eye projection. “We have no room for all these people! We are full!”
Death of exuberance? People of color don’t mess around:
Teenager faces death by firing squad after she was caught with cocaine in her luggage
BRIIC! This involves the two largest nations in the Southern Hemisphere, both of which poke up into the Northern. They are the center of the world.
But here's the problem. The entire BLM "cops are killing blacks!" fraud is debunked by the single statistic that blacks are about 30% of police killings, but account (at the time) for 55% of homicides. (Homicide being a pretty good metric for the sort of situation "guy waving a gun around" call that gets cops showing up with weapons drawn and generates police shootings.) So police if anything are "under shooting" blacks relative to miscreants from other races. But that very clear stat, had zero traction.This simply is not about numbers--math, logic, reason. This is about this now dominant verbalist overclass religious ideology--minoritarianism--virtuous minorities oppressed by big bad racist white gentiles. With its sacred stories (the holocaust, etc.) of which the black experience in America--slavery, Jim Crow--is critical and foremost in the American rite.You aren't going to "debunk" any part of the narrative, with some time series behavior of blacks getting extra "vibrant". That blacks are oppressed by big bad whitey is a fact--a cornerstone of the massive minoritarian edifice, that justifies so much parasitism and looting and destruction.Replies: @Pixo
Agree that the problem has never been lack of data or how to present it visually. It’s about breaking the taboo about racial genetic inequality.
Nathan Cofnas said it well in a substack I linked to a few weeks ago:
“But wokism is protected by a taboo on recognizing the genetic basis of race differences. From early childhood onwards, we’re taught that taking hereditarianism about group differences seriously is a heinous moral crime. All psychologically normal people have been affected by this experience. Until recently, there was a similar taboo on questioning the Biblical account of creation. Darwin himself, when sharing his view that species are not immutable, wrote that it felt like he was “confessing a murder.” In 2023, taboo violators face not just internal psychological resistance, but a very real threat of punishment.
Again, people have to make a decision: we can either live with wokism or do what is necessary to stop it. Whoever wills the demise of wokism wills knowledge of race differences.”
–
It is easier to diagnose the problem than solve it. My best solution is to break the taboo as non-threateningly as possible, but without apology. Steve has done an exceptionally good job of this for decades and is a valuable model.
And the part you won’t like: picking your battles wisely, to the point of complete surrender on the JQ, and switching over to effusive support for Israel and adopting Christian Zionist language like “Judeo-Christian American values.” No better friend, no worse enemy.
It's saying much the same thing. Are all 1100 similar?That "Machiavelli Meets FWP" thing is darkly amusing.But why do you often say "they" instead of "we"??Replies: @Pixo
You're half right.Replies: @Bill Jones
https://i.ibb.co/zh6Fnzw/Screen-Shot-2023-04-24-at-5-34-11-AM.pngYour last chart is clearest, so perhaps this effect can be transposed there.Past my bedtime now though ;)Replies: @Hypnotoad666
Steve — This chart is simply not good. It’s got way too many colored lines and it doesn’t show anything visually dramatic at all. Just a bunch of lines increasing smoothly and steadily over the course of the year — which is exactly what one would expect from graphing the cumulative total of a marginal weekly variable. I think that particular dog won’t hunt.
Basically, you have a product — the “Deaths of Exuberance” meme — that you need to market to other online pundits to get them to spread it while giving you attribution for the coinage. To get your target audience to bite you have to convince them that your intellectual consumer product is: (a) surprising; (b) insightful; and (c) important.
Your marketing problem, however, is that the meme is a composite of at least two very different things — black homicide rates and black driving fatalities. So it’s a complicated bank shot to prove that both statistics are rising for the same reason, and then also prove that the common reason is George Floyd.
Also, when considered all by itself, the idea that de-policing results in more black homicides is not particularly surprising or novel — that’s basically the already much-discussed “Ferguson Effect.” So maybe you should market your meme as the traffic fatality corollary to the Ferguson Effect.
But IMHO, you still need some causation evidence to make the whole thesis stick together. For example, one question I would ask is why are these deaths are still continuing to rise. If they are being cased by a mass black psychological response to the Saint Floyd hysteria, then shouldn’t the effects be dampening over time as the Summer of Floyd recedes into the past? How do we know Floyd-ism has anything to do with it all, rather than some other complicated combination of social factors playing out over time, or hand guns, gas prices or sunspots?
For example, does the increase in black homicides inversely correlate with the number of incarcerations of blacks for homicide? And does the increase in traffic fatalities likewise inversely correlate with traffic tickets written to blacks? A graph that shows such common correlation of homicides/traffic deaths to concrete metrics of de-policing metrics might be compelling to prove they are both symptoms of de-policing.
With a little more causation evidence you might have something more viral — as in, “Hey look de-policing actually has consequences (duh!), and this can be demonstrated by showing how it correlates with two different sets of bad outcomes in two very different domains.”
Should I give it a similar treatment? I definitely agree that clarity and concision are critical to getting a point across to those few who are still open-minded. I'll try some lighter shading and get back in a little while.
I like and am used to graphs in this style and it would be fine in an academic journal, but most people would be confused.
Considerations to perhaps make the two cumulative death graphs (first two in the article) even better:
1. Stack the year legend in reverse order to what you have, ie stack with most recent year on top: this would line up better with how they stack as data.
2. Color spectrum for the different years, eg all purple-blue say, from deep purple to light blue, or similar. Fact that it’s a spectrum stresses relationship and most importantly, progression through the years. Needs to be wide enough spectrum to readily recognize different years though.
3. Completely different colors from above spectrum for the two vertical lines.
The last 3 are the best
Also, when considered all by itself, the idea that de-policing results in more black homicides is not particularly surprising or novel -- that's basically the already much-discussed "Ferguson Effect." So maybe you should market your meme as the traffic fatality corollary to the Ferguson Effect. But IMHO, you still need some causation evidence to make the whole thesis stick together. For example, one question I would ask is why are these deaths are still continuing to rise. If they are being cased by a mass black psychological response to the Saint Floyd hysteria, then shouldn't the effects be dampening over time as the Summer of Floyd recedes into the past? How do we know Floyd-ism has anything to do with it all, rather than some other complicated combination of social factors playing out over time, or hand guns, gas prices or sunspots?For example, does the increase in black homicides inversely correlate with the number of incarcerations of blacks for homicide? And does the increase in traffic fatalities likewise inversely correlate with traffic tickets written to blacks? A graph that shows such common correlation of homicides/traffic deaths to concrete metrics of de-policing metrics might be compelling to prove they are both symptoms of de-policing. With a little more causation evidence you might have something more viral -- as in, "Hey look de-policing actually has consequences (duh!), and this can be demonstrated by showing how it correlates with two different sets of bad outcomes in two very different domains."Replies: @anon, @Mike Conrad, @epebble, @Pixo
It’s all in the bend. Only one line (year) bends much more than the others, and it’s not at the COVID point. Maybe Steve should change the vertical scale from linear to log, along with making the graph taller and narrower, changes that should emphasize the bend more.
Nathan Cofnas said it well in a substack I linked to a few weeks ago:
“But wokism is protected by a taboo on recognizing the genetic basis of race differences. From early childhood onwards, we’re taught that taking hereditarianism about group differences seriously is a heinous moral crime. All psychologically normal people have been affected by this experience. Until recently, there was a similar taboo on questioning the Biblical account of creation. Darwin himself, when sharing his view that species are not immutable, wrote that it felt like he was “confessing a murder.” In 2023, taboo violators face not just internal psychological resistance, but a very real threat of punishment.
Again, people have to make a decision: we can either live with wokism or do what is necessary to stop it. Whoever wills the demise of wokism wills knowledge of race differences.”
-
It is easier to diagnose the problem than solve it. My best solution is to break the taboo as non-threateningly as possible, but without apology. Steve has done an exceptionally good job of this for decades and is a valuable model.
And the part you won’t like: picking your battles wisely, to the point of complete surrender on the JQ, and switching over to effusive support for Israel and adopting Christian Zionist language like “Judeo-Christian American values.” No better friend, no worse enemy.Replies: @Mike Conrad, @JimDandy
For giggles, I just looked up your first post on unz.com.
It’s saying much the same thing. Are all 1100 similar?
That “Machiavelli Meets FWP” thing is darkly amusing.
But why do you often say “they” instead of “we”??
Also, when considered all by itself, the idea that de-policing results in more black homicides is not particularly surprising or novel -- that's basically the already much-discussed "Ferguson Effect." So maybe you should market your meme as the traffic fatality corollary to the Ferguson Effect. But IMHO, you still need some causation evidence to make the whole thesis stick together. For example, one question I would ask is why are these deaths are still continuing to rise. If they are being cased by a mass black psychological response to the Saint Floyd hysteria, then shouldn't the effects be dampening over time as the Summer of Floyd recedes into the past? How do we know Floyd-ism has anything to do with it all, rather than some other complicated combination of social factors playing out over time, or hand guns, gas prices or sunspots?For example, does the increase in black homicides inversely correlate with the number of incarcerations of blacks for homicide? And does the increase in traffic fatalities likewise inversely correlate with traffic tickets written to blacks? A graph that shows such common correlation of homicides/traffic deaths to concrete metrics of de-policing metrics might be compelling to prove they are both symptoms of de-policing. With a little more causation evidence you might have something more viral -- as in, "Hey look de-policing actually has consequences (duh!), and this can be demonstrated by showing how it correlates with two different sets of bad outcomes in two very different domains."Replies: @anon, @Mike Conrad, @epebble, @Pixo
True, yes, but this is why I (and several others here) said that Steve’s last chart is clearer.
Should I give it a similar treatment? I definitely agree that clarity and concision are critical to getting a point across to those few who are still open-minded. I’ll try some lighter shading and get back in a little while.
The most important graph for America.
We are losing a million whites per year and adding 2 million non-whites each year.
The spike is one month long, though. Thus, even the additional 200? 300? deaths of black people from 9/11 will be noticeable.
The perfect chart will not persuade the progressives…they may use your data as evidence that systemic racism is the cause of excess black deaths. Any disparity in outcomes is just further evidence of racism. They control the media, the churches and institutions and our schools. Another chart demonstrating racial disparity will just be ignored or used to demonstrate how racism is the root cause of the disparity. No other explanation will be allowed. Wokism is not based on logic, it is a religion. They gain followers, not via logic, but through their steadfast attack on whites.
I particularly like the one that has date bars that show the beginning of Covid in early March 2020 and then Floyd’s OD with a second bar.Replies: @AnotherDad, @Bill Jones
The graphs tell the tale very well.
The problem is the tale is being told primarily to those to whom they are not a surprise.
I gather that the current mode of effective messaging involves tattoos and I’m not sure Mrs Steve is up for that.
Also, when considered all by itself, the idea that de-policing results in more black homicides is not particularly surprising or novel -- that's basically the already much-discussed "Ferguson Effect." So maybe you should market your meme as the traffic fatality corollary to the Ferguson Effect. But IMHO, you still need some causation evidence to make the whole thesis stick together. For example, one question I would ask is why are these deaths are still continuing to rise. If they are being cased by a mass black psychological response to the Saint Floyd hysteria, then shouldn't the effects be dampening over time as the Summer of Floyd recedes into the past? How do we know Floyd-ism has anything to do with it all, rather than some other complicated combination of social factors playing out over time, or hand guns, gas prices or sunspots?For example, does the increase in black homicides inversely correlate with the number of incarcerations of blacks for homicide? And does the increase in traffic fatalities likewise inversely correlate with traffic tickets written to blacks? A graph that shows such common correlation of homicides/traffic deaths to concrete metrics of de-policing metrics might be compelling to prove they are both symptoms of de-policing. With a little more causation evidence you might have something more viral -- as in, "Hey look de-policing actually has consequences (duh!), and this can be demonstrated by showing how it correlates with two different sets of bad outcomes in two very different domains."Replies: @anon, @Mike Conrad, @epebble, @Pixo
Also, searching for Floyd-Ferguson in 2020-2021 data is really searching for needle in a scrap yard since so many things went wrong those years. If the curves stay good over 2022-2023, there is some value in the hypothesis. But the line for 2022 is already sloping below 2021. So, maybe there will be nothing significant after 2023.
Also, when considered all by itself, the idea that de-policing results in more black homicides is not particularly surprising or novel -- that's basically the already much-discussed "Ferguson Effect." So maybe you should market your meme as the traffic fatality corollary to the Ferguson Effect. But IMHO, you still need some causation evidence to make the whole thesis stick together. For example, one question I would ask is why are these deaths are still continuing to rise. If they are being cased by a mass black psychological response to the Saint Floyd hysteria, then shouldn't the effects be dampening over time as the Summer of Floyd recedes into the past? How do we know Floyd-ism has anything to do with it all, rather than some other complicated combination of social factors playing out over time, or hand guns, gas prices or sunspots?For example, does the increase in black homicides inversely correlate with the number of incarcerations of blacks for homicide? And does the increase in traffic fatalities likewise inversely correlate with traffic tickets written to blacks? A graph that shows such common correlation of homicides/traffic deaths to concrete metrics of de-policing metrics might be compelling to prove they are both symptoms of de-policing. With a little more causation evidence you might have something more viral -- as in, "Hey look de-policing actually has consequences (duh!), and this can be demonstrated by showing how it correlates with two different sets of bad outcomes in two very different domains."Replies: @anon, @Mike Conrad, @epebble, @Pixo
Agree.
I like and am used to graphs in this style and it would be fine in an academic journal, but most people would be confused.
Hmm… is there something that suddenly emerged to get him fired? Oh, the Dominion lawsuit that cost Fox almost 800 million. Tucker knew Sidney Powell was batshit crazy from the start and lied about it. Plus Smartmatic has said they will not go the Dominion route but instead make Fox anchors and executives testify under oath. The discovery process of the lawsuit did Tucker in. He said one thing on air and said another thing in private. His problem is that he is actually smart. He knew the election lies were bs. Maria Bartiromo is actually in a safer position because she is stupid enough to be a true believer. She believed she was telling the truth unlike Tucker
No. The white spike accounts for the overwhelming number of people.
The black spike move is small from 27ish to 30ish. 3 extra deaths per 100K blacks.
White the Hispanic spike jumps from 9ish to around 17ish. 8 extra deaths per 100K Hispanics.
And the white spike is a jump from around 3 to 16ish. That’s 13 points–compared to 3–and the white population is almost fives times bigger than blacks. So–ballpark–13/3 * 5 roughly 8 times as many whites killed on 911 as blacks. (More or less what you’d expect.)
You know, I’m just going to say this, or maybe somehow just deposit this, because it may not have any sort of direct relevance but I simply think it’s worth saying……
When I was a kid, I sort of grew up reading the mad, critical rantings of His Holiness Lester Bangs in CREEM Magazine, and that lunatic, God bless him, sort of taught me how to have a critical and somewhat default-caustic approach to whatever I was reading or absorbing at the time: the Ramones, the NYT accursed op-ed page, Andrew Sarris, Fukuzawa Yukichi, Mark Alan Stamaty, David Byrne, Oe Kenzaburo, Ronald Reagan, whatever.
And I felt that I had a kind of fiduciary obligation to pass that along to any young weirdos who felt they were being left out of the conversation, (I can hear you silent teenage nutjobs in Darwin and Montevideo) and who could see right in front of their faces the non-stop violation of what Confucius calls “the Rectification of Names”: people outrageously giving names to things which are transparently not true. And thus are in violation of the right to Names. In our times, things like “Our Democracy” and “systemic racism” and “whiteness” can suffice for examples of pure malediction, an entirely willful agency to poison the wells of public discourse. Lester HATED that, and THAT is why he thought so highly of say, Johnny Ramone or Iggy Pop (Jim Osterberg). And that is why if you have any opportunity, you should re-read Lester Bangs. And if you can’t do that, you pissed-off young people, that is EXACTLY why I am here cluttering up the bargain bin.
Now go look up who Robert Wilson and Patti are, you dumb rapscallions, you. And get off my lawn!!!
Homicide and motor vehicle death rates seem to start climbing in 2016, Obama’s last full year as President.
Policing resources were diverted from patrolling ghettos to patrolling locations considered to be targets for another terrorist attack.
Less police in the hood makes homie all good.
I think that you need a quad plot with two rows and two columns. Each subplot is its own race, so the upper left plot is (for example) Black, the upper right plot is White, and so forth. Maybe call this out with a box.
Each subplot shows the same series of curves. The X-axis runs from January to December. The Y-axis should be per Capita murders by week or month. You want to show rate of change. Cumulative curves wash that out.
Plot one line in each subplot that demarcates a year. Mark the year on the line somewhere either with a squiggle or a break or a call-out. Don’t use a legend in the subplot (too much data). Use some sort of sequential coloring as the lines advance from one year to the next. Viridus is a good one since it works for the colorblind and is relatively nice looking. For 2020 and later, increase the line strength. Mark on 2020 the relevant dates.
I’m not sure if the Y-axis values should be all scaled the same or not. You would need to make a judgement call after seeing the data both ways.
You probably need better plotting software than Excel. Matplotlib is good if you know Python and can invest time in your plots. If not, Veusz is pretty good as well. It’s also free. I haven’t used it in a while, so I can’t speak to it.
Nathan Cofnas said it well in a substack I linked to a few weeks ago:
“But wokism is protected by a taboo on recognizing the genetic basis of race differences. From early childhood onwards, we’re taught that taking hereditarianism about group differences seriously is a heinous moral crime. All psychologically normal people have been affected by this experience. Until recently, there was a similar taboo on questioning the Biblical account of creation. Darwin himself, when sharing his view that species are not immutable, wrote that it felt like he was “confessing a murder.” In 2023, taboo violators face not just internal psychological resistance, but a very real threat of punishment.
Again, people have to make a decision: we can either live with wokism or do what is necessary to stop it. Whoever wills the demise of wokism wills knowledge of race differences.”
-
It is easier to diagnose the problem than solve it. My best solution is to break the taboo as non-threateningly as possible, but without apology. Steve has done an exceptionally good job of this for decades and is a valuable model.
And the part you won’t like: picking your battles wisely, to the point of complete surrender on the JQ, and switching over to effusive support for Israel and adopting Christian Zionist language like “Judeo-Christian American values.” No better friend, no worse enemy.Replies: @Mike Conrad, @JimDandy
No better friend, no worse enemy.
You’re half right.
You're half right.Replies: @Bill Jones
No.
Here’s something that could do with graphing. A Study of Studies.
Perhaps it could be called Deaths of Cowardice.
Deaths of Obedience?
Deaths of Craveness?
https://www.zerohedge.com/medical/explosive-new-study-finds-face-masks-may-increase-stillbirths-testicular-dysfunction
And the “May have” is far too limp wristed.
2. The investigators do not appear to have separated effects of asymptomatic Covid-19 from purported CO2 poisoning.
The cumulative graphs are too highbrow for a meme, the homicide graphs are too noisy (maybe could use larger bins to emphasize trends). The last one is the best but needs a cleaner x axis (no week labels, just the years).
I’m sorry, is there a worse enemy than the attackers of The U.S.S. Liberty, etc.?
My argument was the half.
Pixo's original statement was a straight inversion of the truth.
There are no "friends" only hostages.
I’ll add this link where I tried to do something like this for Baltimore post-Freddie Gray.
http://mistybeach.com/mark/BaltimoreHomicides.html
I show ‘difference from baseline’ per month. I think it works. I don’t know that it is better than what you have.
I perhaps should have been clearer.
My argument was the half.
Pixo’s original statement was a straight inversion of the truth.
There are no “friends” only hostages.
Deaths of Obedience?
Deaths of Craveness?https://www.zerohedge.com/medical/explosive-new-study-finds-face-masks-may-increase-stillbirths-testicular-dysfunctionAnd the "May have" is far too limp wristed.Replies: @epebble
1. This study was done on mice and rats, not humans.
2. The investigators do not appear to have separated effects of asymptomatic Covid-19 from purported CO2 poisoning.
It's saying much the same thing. Are all 1100 similar?That "Machiavelli Meets FWP" thing is darkly amusing.But why do you often say "they" instead of "we"??Replies: @Pixo
Do you see a flaw in the argument?
Yes I have made it before. Some of the WN complaints about Jews are valid, though typically exaggerated. I think that debate has become tiresome and I don’t really have much to add to it, other than noting Western Jews are suffering the same basic demographic collapse as other white groups.
But the WNs seem to assume without argument that vast Jewish talent, influence, wealth and power might be productively opposed. That seems wrong, and the better tactic is accommodation, interbreeding and a cheerful subservience to the resulting hybrid-invigorated elite. This would be an interesting debate if anyone would take the other side.
“My discovery a couple of years ago of how Deaths of Exuberance (murders and car crashes) respond to headline events like Ferguson and Floyd”
Yes, a correlation. But that does not mean causation, as has been correctly noted before.
You say you loves data. Put your pattern recognition skills to the test to bolster your case.
https://elephrame.com/textbook/BL
Mine the data. Clearly define terms and criteria. Establish how many BLM protests were violent and murderous, along with how and why each protest met that criteria.
My favorite thing about you is your sense of humor.
http://mistybeach.com/mark/BaltimoreHomicides.html
I show 'difference from baseline' per month. I think it works. I don't know that it is better than what you have.Replies: @Steve Sailer
Thanks.