Legalized judicial racism…

I would guess the key question here is what variables are *not* being considered (not in your example, but in the real life evaluations)? and why?

From the linked research paper:

What are the sum total of omitted control variables that could lead to a larger estimate of unwarranted sentencing disparities? (not a question for you, I just think that it's important to consider for anybody doing the research)

Go GATORS!
,WESGATORS

 

Too much common sense in your post.
You must include white man racist and then you’ll get the ‘likes’.

 

No, the biggest problem is thinking that you would understand basic statistical inference if explained as simply as possible. I should have known the guy who tried to argue that the Law of Large Numbers isn't a real thing wouldn't understand even the most basic statistical inference. Obviously, statistics work by doing this sort of thing hundreds, thousands, or millions of times across the entire data set. The use of two observations was to keep it as simple as possible for you to understand the concept. Apparently, it wasn't enough.

One of (although certainly not your biggest) mistakes is that you think that you can claim what they have without even knowing what they have.

I'd suggest that if you don't want to continue to look foolish on topics like this, forcing you to resort to personal insults, that you should take a basic statistics class and annoy that Professor with your basic mistakes.

 

Keep your personal life out of this.

 

Absolutely. These are the types of issues that researchers spend their careers agonizing about. What variables to include or not include. How to control for unobserved heterogeneity. How to handle the endogeneity of variables and related issues like reverse causality. Generally, if you are looking to understand a particular relationship, it is best to look across the literature, in which researchers utilize different methods, different variables, and different assumptions when running their analysis. That is why you see meta-analyses like the one that I linked, as there will be tons of studies on the same issue due to the need to test the relationship in a variety of different ways.

 

Which only matters in this context if you argue that a particular racial group is more likely to do that. If not, it is random variance, which is not necessary to eliminate entirely.

 

That's actually an interesting example. I do think how a defendant comports himself matters, but is there any reason to think that White men are significantly more likely to show respect for the Court and remorse for their victims compared to Black men who are convicted of substantially similar crimes? Intuitively, that wouldn't make sense to me unless one believes those sorts of traits are somehow racially correlated.

 

It didn't take. Perhaps try again.

Also, that quote is, itself, a lie (as Twain himself attributes the quote to somebody who never said it).

 

Yep, your reaction to basic statistical knowledge. Like I said, a refresher would do you a world of good if you are going to continue to make such claims.

 

Man, I'm trying not to post to any social media other than about cycling, but I'll chime in...

(Full disclosure, UFLawyer: I am a criminologist w/over two decades of experience and several peer reviewed studies on judicial decision-making. Not meant to be a fallacious "appeal to authority" just that you should probably know given this specific topic).

I disagree with several of @UFLawyer 's contention about the "sheer volume of data needed...", not being able to make any assumptions about sentencing (hint: every study makes assumptions, often multiple ones, regardless of discipline), and his questionable understanding between individual & system.

As much as the law might be esoteric in various ways to outsiders, the system itself is not per se, at least not when someone takes the time to get informed. But it is often quite different or more nuanced in operation relative to the way it is portrayed in some media, hollywood or by those w/vested interests in keeping the ugly underbelly hidden.

Same time, the bottom line truth is we can't get inside any judges head about their biases unless they explicitly demonstrate it in some way, but we can detect bias in sentencing based on the patterns of sentencing, controlling for important factors such as public vs private defender, pre-sentence investigation report findings, arrest & imprisonment history, contextual factors etc. This is the essence of systemic bias, i.e. patterns across large numbers of people/decisions.

It's imperfect, probabilistic & persistently w/some degree of uncertainty...but that is the nature of research! Unfortunately, some want to use limitations to quickly dismiss the entire enterprise--seems firstly because they don't like what the research shows, but have no other way of countering it.

Curious, and not defensive, but I really would like to know what other field @UFLawyer thinks would be more equipped to study our cj system than criminologists trained to do research?

My question isn't to suggest that those working in the system are all wrong or don't understand it and that criminologists know it all. But it's been my experience over the past two decades having collaborated w/on research & taught numerous folks who work or worked in the system (including colleagues) that their views become less impulsively defensive and/or protective of said institutions/system when they avail themselves to honestly confronting research. This is of course, not easy to do when (often illogical or illiberal) habits of thinking develop where people are quick to judge other's motives, beliefs, biases but unwilling to interrogate their own.

 

Good to see you, Mutz. Hope you're doing well!

 

Thanks brotha! Likewise!

I am still plugging along...kind of still sorting my life out lol.

 

Or even if it's not the case that whites are significantly more remorseful than blacks-which is highly unlikely to be the case--the most pernicious effect of bias imo is that people can perceive differences that aren't there based on biased perceptions then act upon on that perception. Many ways, none of us are immune from this (generally speaking).

 

Is your suggestion that all stats are bs?

If so, I'd think through this some more because to my mind's eye, your using stats (or the implicit suggestion of them: e.g. "tendency" "higher incidence") in your argument doesn't square w/"lies, damn lies, and statistics."

 

Put it like this: Working with, measuring & interpreting data is freaking hard. Numerous very good reasons to retain some skepticism (of the honest variety) and to hold any inferences with a light grip...or some such.

I like to use the example of data as being a slightly fuzzy or out of focus snapshot. So, e.g. a picture of Wes is not Wes himself, but a representation of Wes. Follow from that and it's easier to keep in mind how there will always be flaws & limitations in data and studies from it and therefore what any one of us can infer from data/study.

Same time, almost goes without saying, it's ridiculously easy overinterpret/mistakenly infer or mistakenly/over generalize, thus abuse.

 

Off the top of my head, here are most of the variables that regularly appear in sentencing studies (not to suggest that all of them appear in all studies):

-private lawyer vs public defender
-demographics (race/age/gender/education/employment status/SES/marital and/or parental status/family factors)
-criminal history/dangerousness
-drug use history
-offense at sentencing (e.g. drug crime, property crime, violent crime or misdemeanor/felony etc.)
-weapons involvement
-contextual factors of a case
-bail
-plea bargain vs trial
-court (location/culture, fed/state etc./sentencing guidelines/courtroom workgroup/pre-sentence investigation

Not to you specifically/only wes, anyone care to take a stab at identifying factors that might conceivably play a significant role in sentencing but aren't mentioned above?

I can't. I question the "belief" there are an infinite number of variables that haven't been studied or identified and tend to think that the existence of such variables is exceedingly small (obviously I can't prove this).