Trump dissolves voter fraud commission

[Amica Nostra 4,803]

BY JORDAN FABIAN AND BRANDON CARTER25,984 TWEET SHARE MORE

President Trump on Wednesday dissolved a controversial commission that was set up to investigate his unfounded claims of widespread voter fraud during the 2016 election.

The order brought an abrupt end to a highly touted commission that Trump created last May.

The White House said Trump decided to disband the Presidential Advisory Commission on Election Integrity because several states failed to hand over voter information.

http://thehill.com/homenews/administration/367343-trump-dissolves-voter-fraud-commission

[Dashinka 24,492]

I thought some professors at Old Dominion found that there was a possibility of nearly 1 million non-citizen votes in the last election?

[Tahoma 2,292]

One of the authors of the Old Dominion study said that Trump misinterpreted the study.

 

Even if Trump's misinterpretation were correct, Clinton still would have won the popular vote by over two million votes.

[bcking 3,157]

5 hours ago, Bill & Katya said:

I thought some professors at Old Dominion found that there was a possibility of nearly 1 million non-citizen votes in the last election?

https://ww2.odu.edu/~jrichman/NonCitizenVote.pdf

 

I found this study for Old Dominion, but it isn't looking at anything from the 2016 election. It uses data from 2008 and 2010.

 

They had 339 and 489 non-citizens (self-identified) from the 2008 and 2010 data respectively (Out of a total of something like ~30,000 in 2008, and 55,000 in 2010). That was about 1% in each year, which they comment is an under-sampling of the general population (Approximately 4% in 2008). Apparently it's like 7% more recently? (Based on good ole Google)

 

Of those 300-500 non-citizens, it was around 50-70 that were "registered to vote" (~20%). There are 5 cases (All from 2008, none in 2010) that were self-reported and verified. 51 cases self-reported (but not necessarily verified) voting in total between the two data sets. (6%, 51/828).

 

So if we use 6% of non-US citizens voting as an estimate and 22.3 million for the number of non-citizens (Again - Just from a quick google) that would put it at 1.338 million. So maybe that's where the number comes from.

 

Though in fairness that is combining two election years. In 2008 the rate was 11%, and then in 2010 it was 3.5% so you could use either one of those numbers. You also have to consider the under-sampling and question whether that biases the study towards a certain group of non-US citizens.

[bcking 3,157]

Here's a good quote from the study:

 

"How many non-citizen votes were likely cast in 2008? Taking the most conservative estimate – those who both said they voted and cast a verified vote – yields a confidence interval based on sampling error between 0.2% and 2.8% for the portion of non-citizens participating in elections. Taking the least conservative measure – at least one indicator showed that the respondent voted – yields an estimate that between 7.9% and 14.7% percent of non-citizens voted in 2008. Since the adult non-citizen population of the United States was roughly 19.4 million (CPS 2011), the number of non-citizen voters (including both uncertainty based on normally distributed sampling error, and the various combinations of verified and reported voting) could range from just over 38,000 at the very minimum to nearly 2.8 million at the maximum." (https://ww2.odu.edu/~jrichman/NonCitizenVote.pdf, page 12)

 

For me the issue of whether or not people lied is kind of a moot point. People can lie on ANY survey question at ANY time. There wasn't any particular reason for them to lie on the survey about the citizenship status, and there was ESPECIALLY no reason for them to lie and claim they voted when they weren't citizens. If you think they would be worried about getting caught, why would they check that they were both a non-citizen AND they voted, when that is clearly illegal? Honesty is something you have to assume whenever you poll, and the best you can do is make sure that you make it clear to the people doing the survey that the answers are anonymous and they are protected etc etc.... I don't have any experience with that outside of healthcare though. It's usually easy for us because we can just say "HIPAA and all that..."

 

The bigger issues are:

 

1. It's 2008 data, and we are talking about 2016

2. The 2010 data was quite significantly smaller (none verified, 3.5% claiming they voted, down from 11.3 in 2008), which indicates pretty significant sampling error (or if the numbers are close to real, that a lot can change from election year to election year). Would 2012 be down even more? Up again? What about 2014, then 2016? We have no idea. 

3. Their two confidence intervals are VERY wide, especially the conservative estimate -  0.2% to 2.8%? Those are an order of magnitude different. Even the less conservative (using unverified votes) gives wide CI's (7.9 to 14.7%). Confidence intervals that wide are generally a sign that you need a larger sample. While their starting population was good (30,000+ in 2008), but the real population they care about is quite small, which makes what they care about a "rare event" and therefore hard to study. They honestly just need bigger samples to narrow those CI's 

4. Their overall question is whether Non-citizens vote, they don't specify a specific election year. And yet in the end as far as I can tell they only provide confidence intervals for the 2008 year. I may have missed their analysis of the 2010 data set. As far as I can tell from their methods they were intended to analyze both the 2008 and 2010 data, so it's not clear why they only fully report one. In scientific literature we call that "reporting bias" because it makes you question why they aren't reporting data that they should have, and that it appears they planned to generate. It makes you wonder if they aren't reporting it because it presents a result that they don't like. 

 

This isn't to say the study is bad. This is just a brief 10 minute critical analysis. I'm critical about everything, even what I publish. I haven't read through the whole thing as well. Just the methods and the relevant results section.

Edited January 4, 2018 by bcking

[Amica Nostra 4,803]

35 minutes ago, bcking said:

Here's a good quote from the study:

 

"How many non-citizen votes were likely cast in 2008? Taking the most conservative estimate – those who both said they voted and cast a verified vote – yields a confidence interval based on sampling error between 0.2% and 2.8% for the portion of non-citizens participating in elections. Taking the least conservative measure – at least one indicator showed that the respondent voted – yields an estimate that between 7.9% and 14.7% percent of non-citizens voted in 2008. Since the adult non-citizen population of the United States was roughly 19.4 million (CPS 2011), the number of non-citizen voters (including both uncertainty based on normally distributed sampling error, and the various combinations of verified and reported voting) could range from just over 38,000 at the very minimum to nearly 2.8 million at the maximum." (https://ww2.odu.edu/~jrichman/NonCitizenVote.pdf, page 12)

 

For me the issue of whether or not people lied is kind of a moot point. People can lie on ANY survey question at ANY time. There wasn't any particular reason for them to lie on the survey about the citizenship status, and there was ESPECIALLY no reason for them to lie and claim they voted when they weren't citizens. If you think they would be worried about getting caught, why would they check that they were both a non-citizen AND they voted, when that is clearly illegal? Honesty is something you have to assume whenever you poll, and the best you can do is make sure that you make it clear to the people doing the survey that the answers are anonymous and they are protected etc etc.... I don't have any experience with that outside of healthcare though. It's usually easy for us because we can just say "HIPAA and all that..."

 

The bigger issues are:

 

1. It's 2008 data, and we are talking about 2016

2. The 2010 data was quite significantly smaller (none verified, 3.5% claiming they voted, down from 11.3 in 2008), which indicates pretty significant sampling error (or if the numbers are close to real, that a lot can change from election year to election year). Would 2012 be down even more? Up again? What about 2014, then 2016? We have no idea. 

3. Their two confidence intervals are VERY wide, especially the conservative estimate -  0.2% to 2.8%? Those are an order of magnitude different. Even the less conservative (using unverified votes) gives wide CI's (7.9 to 14.7%). Confidence intervals that wide are generally a sign that you need a larger sample. While their starting population was good (30,000+ in 2008), but the real population they care about is quite small, which makes what they care about a "rare event" and therefore hard to study. They honestly just need bigger samples to narrow those CI's 

4. Their overall question is whether Non-citizens vote, they don't specify a specific election year. And yet in the end as far as I can tell they only provide confidence intervals for the 2008 year. I may have missed their analysis of the 2010 data set. As far as I can tell from their methods they were intended to analyze both the 2008 and 2010 data, so it's not clear why they only fully report one. In scientific literature we call that "reporting bias" because it makes you question why they aren't reporting data that they should have, and that it appears they planned to generate. It makes you wonder if they aren't reporting it because it presents a result that they don't like. 

 

This isn't to say the study is bad. This is just a brief 10 minute critical analysis. I'm critical about everything, even what I publish. I haven't read through the whole thing as well. Just the methods and the relevant results section.

Do these responses include Residents who vote in local elections such as School levies which is legal in some cities?  

[bcking 3,157]

17 minutes ago, Il Mango Dulce said:

Do these responses include Residents who vote in local elections such as School levies which is legal in some cities?  

No it was questions relating to federal offices only I believe. President, and Senate/House.

 

Though you could wonder if some people polled thought the question applied to local votes that are legal for them and maybe got the question confused.

 

They authors noted that Maryland had some situations where voting would be legal, but that none of the people who responded claimed to have voted there.

[bcking 3,157]

I'm reading more of it intermittently throughout the day when I have time in between work. I just happen to enjoy reading these sorts of things. A couple of other things from their methods:

 

1. "Many non-citizens were asked if they voted, unlike other large surveys which filter out non-citizens before asking about voting." - Why just "many"? Why not all? I'm not sure if this is just poor written or if they aren't properly explaining the original dataset. I'm not sure why the survey would only ask some but not others.

 

2. "Participation and registration were verified for at least some residents in nearly every state for the 2008 survey" - Again, why just "some"? How did it choose which ones to verify? Was it a random subset because verification is time consuming? If so, they should specify that it was random (which would be good). If it was not random, then that's a problem.

 

3. The original datasets are weighted, but then the researchers use their own rationale and methodology for weighting compared to census data. Since we are dealing with fairly small numbers (1% of the original surveyed population), even minor adjustments with weighting can produce significant changes and weighting just adds more assumptions which can be good or bad. I think they have non-weighted data in their appendix, but I haven't looked for that yet.

 

Another bit of data as well that is useful from the study:

 

"71 non-citizens answered a survey question indicating whether they voted, and also had their vote validated. Among these, 56 indicated that they did not vote (but two of these cast a validated vote, while 13 indicated they voted, of whom five cast a validated vote."

This goes back to #1 and #2 - Why these 71? Were they are random sampling? They don't explain how it was decided to validate some but not others. That's reporting bias.

 

 

[Voice of Reason 10,298]

3 hours ago, Tahoma said:

One of the authors of the Old Dominion study said that Trump misinterpreted the study.

 

Even if Trump's misinterpretation were correct, Clinton still would have won the popular vote by over two million votes.

I’ll type this slowly in the hopes it helps comprehension set in without too much pain.

 

H I L L A R Y   L O S T   B E C A U S E   S H E   W A S   A   H O R R I B L E   C A N D I D A T E

[bcking 3,157]

Just now, IDWAF said:

I’ll type this slowly in the hopes it helps comprehension set in without too much pain.

 

H I L L A R Y   L O S T   B E C A U S E   S H E   W A S   A   H O R R I B L E   C A N D I D A T E

The quality of her candidacy is not really the issue of this thread, it's a little off topic.

 

We are talking about voter fraud. On that topic Tahoma's point is completely valid. We should be able to talk about these things without people flipping out, or posting in all caps...

 

Clinton gained about 3 million more votes. Even if we take the study's numbers at face value, it would be about 1.3 million non-citizen voters. She would still have won the popular vote by "around" 2 million (More like 1.7 million or so).

[Voice of Reason 10,298]

2 minutes ago, bcking said:

The quality of her candidacy is not really the issue of this thread, it's a little off topic.

 

We are talking about voter fraud. On that topic Tahoma's point is completely valid. We should be able to talk about these things without people flipping out, or posting in all caps...

 

Clinton gained about 3 million more votes. Even if we take the study's numbers at face value, it would be about 1.3 million non-citizen voters. She would still have won the popular vote by "around" 2 million (More like 1.7 million or so).

Welcome to America, where popular votes are cute, but get no one elected.  Even if Hillary had over 1 million illegal votes, she still didn’t win.  Not sure why the popular vote is even being brought up.  Again.

[Amica Nostra 4,803]

4 minutes ago, IDWAF said:

I’ll type this slowly in the hopes it helps comprehension set in without too much pain.

 

H I L L A R Y   L O S T   B E C A U S E   S H E   W A S   A   H O R R I B L E   C A N D I D A T E

As much as some might wake up every day a thank their lucky stars that Hillary is alive to hate and despise, Hillary is not the topic.

Just now, IDWAF said:

Welcome to America, where popular votes are cute, but get no one elected.  Even if Hillary had over 1 million illegal votes, she still didn’t win.  Not sure why the popular vote is even being brought up.  Again.

Because the snowflake president wasted money and time on the topic.

[bcking 3,157]

3 minutes ago, IDWAF said:

Welcome to America, where popular votes are cute, but get no one elected.  Even if Hillary had over 1 million illegal votes, she still didn’t win.  Not sure why the popular vote is even being brought up.  Again.

What's being discussed is the number of non-citizens who may have voted in the election. From the onset of this topic no one here has claimed it would have changed either outcome (Hillary still would have won the popular vote, Trump would still be President). Trump started the commission, and he has dissolved it. The question is about whether or not it would have changed the outcome.

 

Literally the only person here who has even suggested that the outcome would be different is you. You come in here and try to create drama where there is none. We can have a discussion of voter fraud, and the study that has been referenced, without adding conflict. If you don't have anything to contribute to the discussion, go somewhere else and try to create conflict.

Edited January 4, 2018 by bcking

[Voice of Reason 10,298]

4 minutes ago, bcking said:

What's being discussed is the number of non-citizens who may have voted in the election. From the onset of this topic no one here has claimed it would have changed either outcome (Hillary still would have won the popular vote, Trump would still be President). Trump started the commission, and he has dissolved it. The question is about whether or not it would have changed the outcome.

 

Literally the only person here who has even suggested that the outcome would be different is you. You come in here and try to create drama where there is none. We can have a discussion of voter fraud, and the study that has been referenced, without adding conflict. If you don't have anything to contribute to the discussion, go somewhere else and try to create conflict.

There is no voter fraud.  Hasn’t been for 20 years, dontcha know that?  

 

Where’s Teddy when you need him?

[Tahoma 2,292]

2 hours ago, IDWAF said:

I’ll type this slowly in the hopes it helps comprehension set in without too much pain.

 

H I L L A R Y   L O S T   B E C A U S E   S H E   W A S   A   H O R R I B L E   C A N D I D A T E

I'll type this slowly so that you can comprehend without too much pain.

 

YOU ARE OFF TOPIC