Roy Moore accused of sexual contact with 14-year old

[carpedm9587]

OK, I give up, because you set up a nearly impossible "condition", which gives you cover to claim you support Voter ID laws, but.....

You can have the last word.

Nah...

(but I have to admit to morbid curiosity about this "condition" I set up that has given me this cover. I truly have no idea what you're talking about)

[Sparko]

Umm.. go back and read my posts, Sparko. I have been fairly clear that VoterID is not, per se, a problem. Disenfranchisement is. If VoterID requirements are implemented before a VoterID infrastructure is in place and registered voters without IDs are equipped with those IDs (or have been given reasonable opportunity and resources for doing so), then I object on the basis of what I have already outlined. I have voiced admiration for an ID system (see the post about South Africa), and explicitly noted the disconnect with what is being done here.

If I have been sloppy with language, please point to the posts that were problematic and I will clarify or appologize. If we have been agreeing violently, then I certainly apologize. Your posts left me with the impression that going gangbusters with a VoterID program without taking time to ensure no one's vote would be disenfranchised. Indeed - I seem to recall you explicitly saying that you didn't care if some validly registered voters couldn't vote as a result. Didn't you express that opinion?

No I said that to be a validly registered voter there needs to be some criteria, like being a US citizen and a resident of the state and county where your voting district is. That in order to prove your qualifications requiring an ID is a valid thing to ask for. And that we should make it as easy to obtain as it is possible to do. BUT that if there are those out there who cannot meet the criteria for some reason (no residence, no birth certificate, etc) then I am good with some people not being able to vote. There is no way to satisfy everyone all the time. Some people will not qualify. You can't just toss out the requirements to satisfy those few. But we should make it as easy and cheap as possible for those who can meet the requirements to get a valid ID to vote.

This suggests we might be agreeing violently. I think the only possible point of disconnect is that I believe the VoterID system should be implemented and reasonable efforts made to make sure everyone who is a validly registered voter has one BEFORE it is mandated in order to vote. Based of the database of actual voter fraud, the delay will have negligible impact on voting, a valid assumption because there is negligible evidence of voter fraud to date.

The criteria you must meet to vote is already there though: You have to be a citizen, you have to have a valid residence in the district you are voting in, you can't be a felon, and so on. The voter ID just proves that you meet those criteria.

[Teallaura]

No - I'm not. I'm citing the fact that this database does not say what it is being purported to say - and the statistics derived from it are not adequate to substantiate a case for widespread voter fraud at a statistically significant rate. There is some evidence of the cases that WERE brought to trial and a conviction was not possible. There is no data, as far as I know, about htings NOT brought to trial. So I cannot make an argument against the claim that voter fraud is widespread on the basis of that information, and you cannot make a case for a claim that it is either. The absence of evidence is not an argument in this situation.

Um, no - this is the instance in which weighting becomes necessary. You can't ignore a variable simply because you don't have a way to get good numbers on it. Your ratio is garbage if it doesn't include the consideration of the numbers not found in the conviction numbers - which are necessarily skewed low.

As little as I like weighting, it is a necessary evil of statistical analysis - and it's not going to be your friend. Conviction rates of voter fraud - like most conviction rates - will be much lower than the incidence rate. Even guessing conservatively, the numbers are going to be significantly higher than the conviction rate to have any validity. Otherwise, it's a flawed statistical model that has no external validity - which makes the math all pretty but the real world application crap.

[Teallaura]

You and others here are ignoring a reality you just don't seem to want to face. A significant number of people without the proper ID (as per the studies that have been made) also lack the means to get them. They lack the source papers (e.g., birth certificates), they lack the financial means (yes, tens of dollars matters to the poor), they lack the logistic means, etc. None of these things is beyond repair. All of them can be addressed. But it requires organization, putting resources in place, and doing the outreach. Meanwhile, a law that creates a change in voting requirements and effectively adds an obstacle not previously present, without ensuring that these issues have been addressed, creates an obstacle to voting. The studies show that these obstacles are disproportionately impacting the poor - which means disproportionately impacting minorities - which means skewing the vote to the Republican Party. The numbers from the study are 2% with a 1.2% swing of the voting outcomes. Even if the number was 0.2%, an order of magnitude less, the number of people impacted is still more than 3 orders of magnitude (2,000 times) greater than the incidence protected against. And that's being generous with the numbers. If we use the ACTUAL numbers from the studies (2%) and the limited number of fraud types that would be impacted based on the database that was provided), the number is actually closer to 5 orders of magnitude (100,000 times as many).

None of these things is necessary or can be justified. The "voter fraud" the ID's proport to solve cannot be shown to be statistically significant, and the number of people impacted by the new regulations is simply too high to leave unaddressed, especially when there is a simple solution: implement the infrastructure for an ID system, ensure reasonable effort has been made to outreach to all existing reigstered voters and ensure all of those validily registered have a reasonable opportunity to secure the necessary materials and are not denied access to vote, and THEN make the VoterID mandatory. No one is disenfranchised. Concerns about the handful of voter fraud instances that CAN be addressed by IDs is addressed (leaving all the rest, of course).

I'm not sure why this concept is so difficult to grasp, or why this approach is seen as so undesirable. The impression being left is that the real objective of those advocating for this is not actually voter fraud - but rather to achieve a skewing of the vote that would be the effect for a period of time. There is no other explanation that makes sense. The data is clear. It is unambiguous. And a simple solution presents itself to avoid this effect.

*emphasis mine

There are no such studies linked - your assumption about the one you linked is unsubstantiated - disenfranchising fraudulent voters is the point and could just as easily explain the numbers in that study.

[Cow Poke]

Um, no - this is the instance in which weighting becomes necessary. You can't ignore a variable simply because you don't have a way to get good numbers on it. Your ratio is garbage if it doesn't include the consideration of the numbers not found in the conviction numbers - which are necessarily skewed low.

As little as I like weighting, it is a necessary evil of statistical analysis - and it's not going to be your friend. Conviction rates of voter fraud - like most conviction rates - will be much lower than the incidence rate. Even guessing conservatively, the numbers are going to be significantly higher than the conviction rate to have any validity. Otherwise, it's a flawed statistical model that has no external validity - which makes the math all pretty but the real world application crap.

Kinda like the court docket showing people convicted of running a red light does not indicate the number of people who actually run red lights.

But, it's worse in the example of voting, because the rules vary so wildly, the detection is bad, and there's this huge stigma that if you actually question somebody trying to vote - especially a minority [gasp] you're guilty of VOTER SUPPRESSION and/or DISENFRANCHISEMENT.

[carpedm9587]

Um, no - this is the instance in which weighting becomes necessary. You can't ignore a variable simply because you don't have a way to get good numbers on it. Your ratio is garbage if it doesn't include the consideration of the numbers not found in the conviction numbers - which are necessarily skewed low.

As little as I like weighting, it is a necessary evil of statistical analysis - and it's not going to be your friend. Conviction rates of voter fraud - like most conviction rates - will be much lower than the incidence rate. Even guessing conservatively, the numbers are going to be significantly higher than the conviction rate to have any validity. Otherwise, it's a flawed statistical model that has no external validity - which makes the math all pretty but the real world application crap.

Emphasis is mine. Without a basis for the guess - that is all you are doing: guessing. It is not a basis for making or substantiating a claim.

[carpedm9587]

*emphasis mine

There are no such studies linked - your assumption about the one you linked is unsubstantiated - disenfranchising fraudulent voters is the point and could just as easily explain the numbers in that study.

I've linked to this, and other, sources several time. Here it is again: https://www.brennancenter.org/analys...ter-fraud-myth

The articles itself contains links to several of the studies I've referenced.

[rogue06]

You and others here are ignoring a reality you just don't seem to want to face. A significant number of people without the proper ID (as per the studies that have been made) also lack the means to get them. They lack the source papers (e.g., birth certificates), they lack the financial means (yes, tens of dollars matters to the poor), they lack the logistic means, etc.

So how exactly did these people function in a modern society which pretty much requires an ID to do just about anything these days?
Apparently they're shut-ins who have others do everything for them including pay for everything since they don't have checks, credit cards, or even cash never having had a job (which requires an ID to get).

[Teallaura]

Emphasis is mine. Without a basis for the guess - that is all you are doing: guessing. It is not a basis for making or substantiating a claim.

Dude, do you know what weighting is?!?!?! Seriously, stop talking about the statistical analysis of survey research until you look it up.

Your analysis has no external validity without it or some way to get actual numbers.

[Teallaura]

I've linked to this, and other, sources several time. Here it is again: https://www.brennancenter.org/analys...ter-fraud-myth

The articles itself contains links to several of the studies I've referenced.

Not in this thread, I couldn't find it.

Great, a lit review - back in a few days when I've had time to read some of them.

[carpedm9587]

Dude, do you know what weighting is?!?!?! Seriously, stop talking about the statistical analysis of survey research until you look it up.

Your analysis has no external validity without it or some way to get actual numbers.

Yes - I do. And it's mathematically significant. But you've simply pointed to a statistical process - and claimed it lends weight to your argument. You haven't done the math or made the case. Indeed - I have not seen that math or case done anywhere by anyone.

[carpedm9587]

Not in this thread, I couldn't find it.

Great, a lit review - back in a few days when I've had time to read some of them.

The Brennan report links to multiple studies done in the past decade+. Have you looked at ANY of them?

[Teallaura]

Yes - I do. And it's mathematically significant. But you've simply pointed to a statistical process - and claimed it lends weight to your argument. You haven't done the math or made the case. Indeed - I have not seen that math or case done anywhere by anyone.

Nope, I've pointed out twice now that your ratio ignores a major variable - that makes it externally invalid. pretty math but meaningless in the real world.

And if that means your lit review studies are the same, you just refuted them yourself. Bad methodology equals bad results.

[Teallaura]

The Brennan report links to multiple studies done in the past decade+. Have you looked at ANY of them?

What did you think I was referring to? I've seen a refutation that seems to back up your contention that none of the studies allow for the actual incidence - if so, you've linked to a lot of bad studies. I haven't had time to review them - as I stated.

But I DID see your ratio and again, it is missing a major variable and is externally invalid.

[carpedm9587]

Nope, I've pointed out twice now that your ratio ignores a major variable - that makes it externally invalid. pretty math but meaningless in the real world.

And if that means your lit review studies are the same, you just refuted them yourself. Bad methodology equals bad results.

I did not use the studies for my math - my math was simply pointing out the incorrect use being made of the database cited earlier - and the math was solid. The point was, without the weighting, you have a 10,000:1 ratio - and that's on the generous side (i.e., including all of the forms of fraud voter ID would NOT address). You would have to make the case that for every indicted case of voter fraud, over 10,000 cases of non-indicted voter fraud exist - just to break even.

Those studies show that the projected number of such cases doesn't even begin to approach those values.

There's nothing wrong with my math, Teal. If there is - then you need to show it. And dismissing over a dozen studies as "bad" without even reading them is not going to cut it. I'm aware that my arguments and the math ar enot going to convince anyone here. We've reached a point in our society there data and research have simply become irrelevant. Heck, even the CDC has recently been told by the executive branch that they may not use the terms "evidence-based" and "science-based" in their documentation (specifically budget request, I believe). But they are not irrelevant to me - they serve as the basis for my conclusions.