This is how you see it. Your first response being "How is this study wrong?" rather than "How might my assumptions be wrong?" kind of proves the point.
Not at all. My first question was which of my assumptions are wrong, then double check that my facts were right. Those didn't line up with the article, which is when I stated to see interesting unexplained choices, like using Trump instead of governor or a matrix of policies.
And if they used Governor, you might complain that they didn't use Statehouse. And if they used that, you might come up with another reason.
My guess is that they used Trump AND Biden because they were the only politicians on the ballot in each of the states they studied. Many states didn't even have a Governor race in 2020 (the year they studied for election and murder).
The best thing would likely be to look at a matrix of policies for each state to identify potential impact from a specific policy over time and in combination with the other factors.
As I mentioned, likely because it was a data point that crossed city, county, and state lines. I can't think of another data point that provides that type of insight across such a broad swath of the population at a specific point in time. If you have some, please provide, it would make for a richer data set.
> The best thing would likely be to look at a matrix of policies for each state to identify potential impact from a specific policy over time and in combination with the other factors.
Plenty of that analysis is available, unfortunately the results don't penetrate the echo chamber. Here are a few [1] [2] [3].
[1] https://www.ojp.gov/pdffiles1/nij/grants/254669.pdf
[2] https://www.rand.org/research/gun-policy/analysis/stand-your...
[3] https://www.bu.edu/articles/2019/state-gun-laws-that-reduce-...
That's largely not what we want when looking at policy that is typically enacted at the local level.
I asked for a matrix analysis with how various policies interact and perform over time, not a single subtopic. Those seem to be on a subtopic, and generally not related to the policies typically cited for being soft on crime (topic at hand). They could be part of the matrix as they are indirectly connected.
Are you aware that two of your sources contradict each other? One claims that stand your ground laws have no impact and another claims they increase violent crime and firearm homicide. Also interesting is that Rand omits some studies entirely. I would have liked to see what flaws they would have found in them rather than just leave them out (and leave us wondering if it was overlooked or due to a flaw).
The power of the governor varies substantially from state to state.
> plus they are the ones who set regulations (although city level regulations can vary wildly)
At least in most states, the laws are crafted by the legislature, and implemented by the Governors. Either way, I think it's very reasonable to assume that the in states that voted for a particular President, the local regulations within that state are more likely to be in line with that President and their party than if they voted the other way.
> The best thing would likely be to look at a matrix of policies for each state to identify potential impact from a specific policy over time and in combination with the other factors.
Yes. I would imagine if such a matrix showed that the policies of the Democratic party, normalized for other factors, were correlated with higher murder rates, the prevailing narrative would be well founded.
The point is, absent that, there's no reason to accept or speak to that narrative, other than to mislead and disseminate disinformation for political gain.
I think this is our main issue - which interpretation of the narrative are we discussing. A lot of times each side takes the worst interpretation or intention of the other. In the HN charitable spirit, I interpret the Republican narrative as questioning if policies like doing away with pretextual stops or sanctuary cities contribute to increases in crime. Likewise I interpret this article as trying to support those policies by showing there's little difference between the states. The part I take issue with, is that just analyzing based on partisan status and not on policies missed the debate around policies and doesn't control for socioeconomic factors. If they had focused on those, then it could provide much stronger evidence that the policies in question aren't contributing.
Now maybe I have it all wrong and this article is simply bashing Republicans (some of the text suggests that)
"Either way, I think it's very reasonable to assume that the in states that voted for a particular President, the local regulations within that state are more likely to be in line with that President and their party than if they voted the other way."
Potentially. In my state the legislature frequently passes things that are vetoed by the governor. The majority of the population is also in cities aligned to the governor's party (eg policies that start in the cities are requested by and sometimes passed on their request). I'm sure it's different in other states too. One big factor too, is that there was a lot of hate for Trump that could have nudged some middle of the road states away from their typical voting or the governor party(NH, VT, RI etc) which would affect the averages by party.
But you are sort of right - local policy tends to be the ones discussed. So even in "red" states, some of the large cities could be implementing "blue" policies. Although the same can be true the other direction - not all cities in "blue" states are implementing the policies in question.
So, for starters "sanctuary cities" is not a policy, but a narrative. There is no legal definition for "sanctuary city", and it was a term that was crafted for the narrative about seven years ago, long after some cities made choices to limit the sharing of information between their local law enforcement and the federal government's immigration enforcement in order to establish trust with those communities, and terms like "City of Refuge" that had historically been adopted, but the term implies an entirely different context. The policies weren't really partisan when they were enacted. They were more about law enforcement and community relations, and were primarily advocated for by religious, rather than political, groups. The practical challenges they were responding too are just disproportionately likely presenting in cities that broadly vote Democrat, so the narrative that they are partisan has itself has been constructed. Also worth noting: the broader Republican narrative is for local government to be independent and not work closely with national government or law enforcement, so this peculiar exception is noteworthy and hard not to attribute to deliberate propaganda rather than specific policies.
So just with that statement there's a lot of partisan context you're bringing to the table there that isn't in the article.
> Likewise I interpret this article as trying to support those policies by showing there's little difference between the states.
I totally agree it makes no substantive claims about those policies one way or the other. As you've pointed out it cites no data tied to policies whatsoever. It looks at changes in 2020, and as far as I know, there were no cities that changed their policies around immigration in 2020. It doesn't actually reference immigration or sanctuary cities. It references "liberal cities" in its conclusion, and only in reference to the unsupported narrative. The closest it comes to referencing policy at all is pointing out that the high-murder rates are states that haven't "even flirted with ideas like defund the police", a claim which doesn't even imply that "flirting with", let alone implementing, a policy would have any kind of impact.
Interpreting it as supporting "sanctuary cities" seems as reasonable as claiming it is supporting policies of higher taxes, abortion, or really any policy prescription at all. The one policy it references, "defund the police", which at least sort of fits the comparative timeline of the data it is looking at, is referenced in the article in passing in the conclusion, not anywhere in the establishing argument, and even the reference is about the lack of evidence of a link, rather than there being a positive association. Even concluding the article is supporting the idea that simply voting Democrat would help seems like a huge stretch. I would suggest that seeing it as being about "sanctuary cities" is bringing a lot of your own context to the article that isn't from the article itself.
I would agree that if the aim is to support such policies, and in fact the only reference it makes to any policy at all is a passing reference in the conclusion where it asserts the policy is NOT associated with the data, it does indeed do a disastrous job what with the not mentioning them or providing any data tied to them. It's hard to imagine that someone intending to support such policies with an article would craft such a thing.
Interestingly, I do think it supports a perfectly rational reason why Republicans might be disproportionately concerned about an increase in murder rates, because they are disproportionately increasing in the areas where they live. ;-)
Let's look at the claims in the conclusion: "But the data clearly paint a different story. The increase in murders is not a liberal cities problem but a national problem. Murder rates are actually higher in Republican, Trump-voting states that haven’t even flirted with ideas like defund the police. Eight of the ten most lawless, high-murder states are not only Trump-voting states, but GOP bastions for the last quarter of a century. A more accurate conclusion from the data is that Republicans do a far better job blaming others for high murder rates than actually reducing high murder rates."
It makes exactly zero statements about Democrat policies. It does speak to Republican support and the dearth of Democrat policies, but that is what you'd expect if the intent is to dispute a claim that the Democrat policies are causal or at least correlated with murder rates. It claims the problem is national which doesn't align at all with the interpretation that it is supporting those policies other than perhaps indirectly by saying, "there's a national problem, and the claim that it's due to local policies doesn't align with the trends in the data".
I think one could reasonably interpret it as critical of Republicans, the Republican party, or at least a narrative which it associates with Republicans, but that is a far cry from supporting a policy of any kind. I'd be inclined to agree with you, though I'd argue the core thrust of the article is just that the narrative is not well supported.
That's the conceit of the partisan divide: that anything critical of your tribe, or even anything associated with your tribe is intrinsically supportive of the "other" tribe or anything associated with the "other" tribe. It's unsurprising that one might fall victim to the conceit, but I hope you can recognize that's exactly what has transpired here.
"Potentially. In my state the legislature frequently passes things that are vetoed by the governor."
That narrows things down to maybe 49 of the 50 states. ;-)
"One big factor too, is that there was a lot of hate for Trump that could have nudged some middle of the road states away from their typical voting or the governor party(NH, VT, RI etc) which would affect the averages by party."
That is very much true. However, as the article points out, to the extent any particular state is disproportionately impacted by the trend, none of them are none of those states. As the article points out, the top states for the trend aren't even mentioned in the narrative, and aside from #2 Louisiana, the top 5 are broadly Republican strongholds. Of the top 10, you could say Louisiana, New Mexico and maybe Georgia are swing states (though IIRC the Georgia legislature and governor's office haven't been controlled by the Democrats for nearly a generation).
> So even in "red" states, some of the large cities could be implementing "blue" policies. Although the same can be true the other direction - not all cities in "blue" states are implementing the policies in question.
Yes, and this is why the narrative is effective despite being totally misleading. The loathing of "blue" policies within "red" states is disproportionately higher (for many of the same reasons that people are much more inclined to believe the 2020 election was "stolen" if they were red voters who lived in or nearby a blue county), who are consequently much more vulnerable to the confirmation bias the narrative relies on.
It's hard to control for differences between the contexts of cities, but if the local policies were driving the trend, you would expect that states whose population is most significantly living with those policies, particularly the states mentioned in the narrative: California, New York, and Illinois, you would expect they would be the ones with the higher murder rates. Illinois's 9.2 is the closest to something supporting that, with California & New York being down at 5.59 and 4.11 (particularly weird since even compared to Illinois they have far more of their population living in these "blue cities" with anything resembling these policies).
So it's a good narrative from the perspective of political objectives, but it's absolutely terrible from a policy & governance perspective.
"It makes exactly zero statements about Democrat policies. It does speak to Republican support and the dearth of Democrat policies, but that is what you'd expect if the intent is to dispute a claim that the Democrat policies are causal or at least correlated with murder rates."
I would agree, except you need to look at the mitigating and confounding factors to draw that conclusion - something this article doesn't do. This is the big point in my positions (just that the article is lacking, engaging in similar foul play, and doesn't present the level of evidence that most people are claiming). There can be disproportionate rates of various socioeconomic factors that's are not controlled for here which could skew the data. Basic things like not accounting for the other variables responsible make it impossible to assign any value to the one being scrutinized, not looking for random errors (eg Gaussian distribution). Along this same vein, there's no explaination of the outliers. This is a political article, not a well formed study.
It seems that we aren't making any progress here, and my comments are becoming more and more misinterpreted. I guess I'm out.
You interpreted the article as "trying to support those policies, by showing there's little difference between the states". Except it is pointing out that there are significant differences between the states, and they don't align at all with the narrative.
I am pointing out that your interpretation has imbued the article with an objective that by your own description does not correlate with the contents of the article. You've effectively dismissed the article by establishing new goals and goalposts and declaring it a failure.
"I would agree, except you need to look at the mitigating and confounding factors to draw that conclusion"
You are presuming that the conclusion is "these policies are not causing increased murder rates", despite me pointing out that wasn't the objective and both of us failing to finding any argument supporting that conclusion in the article.
> This is a political article, not a well formed study.
It's definitely not a study, and yes it is a political article. Where you lose me is saying this, seemingly acknowledging that the existing narrative is political and not a study... and then saying the article is a failure because it failed to provide what you'd want in a well formed study of the underlying cause.