Debunk "one in 77 billion chance Hillary won the primary without voter fraud"

Politics & Elections
21

Its hard to debunk this because there is so little detail. The only actual reference I could find was to the Stanford paper which was written by two graduate students with no indication of what department they come from. Hopefully, for the sake of their future careers, they aren’t from the statistics department.

There are a couple of things that look fishy. First they only looked at a subset of states (which didn’t include Wisconsin First for some reason they chose to analyze the number of delegates assigned rather than the number of votes.

This makes no sense, since what they are claiming is evidence of vote number fraud it would the potentially corrupted source is exactly what they would want to analyze. Also using delegates is going to add additional biases due to differences in methods used to assign delegates (fortunately it looks like they didn’t count super deligates which would entirely invalidate the analysis.) The only reason I could think of that they wouldn’t do this is if the analysis using actual vote totals didn’t give the results they wanted. Even then all that they show is that those with paper trails are more likely to go for Sanders. But it may be that those states whose are more on the left radical bent are more likely to elect legislators who feel the need for paper trails.

Their next argument is that since caucuses are more public and more tamper proof, the fact that Sanders won more of those is further evidence of fraud. This is clearly spurious since it is well known that the population that votes in a primary is very different from the population that votes in a Caucus.

They also make some statements regarding the differences between Exit polls and final results. The data suggest a fairly consistent bias of about 3%, with one big exception in Arizona (a paper trail state ;)) in which there was a 20 point bias. Overall its not surprising that there would be a small consistent bias for the reasons Ulf outlined.

Finally they say that the results they got didn’t happen in the 2008 primary, so there must be something weird about this election specifically. This isn’t exactly surprising, since that was an entirely different election with an entirely different set of issues that could create sampling biases.

The plots regarding the black voting patterns are shown without any context or statistical analysis at all, so I don’t know what to make of them. If anything I would think that they actually work against the claims made. Since the only way to find out the % of blacks who voted for Clinton would be through exit polls which wouldn’t be subject to voter fraud. Indicating that the difference they saw between such states was not related to fraud.

Next set of “evidence” is that Sanders has higher approval ratings than Clinton. Especially among Millennials. Well, so what, we knew that. Lots of voters don’t particularly like Clinton as a person all that much but think she is a better candidate. And lots of Millennials say they are going to vote but don’t actually show up.

Finally they have some plots, relating time at which vote totals came in and size of precinct in some cumulative way that seems like a very odd way to do the analysis. Again, I assume that they didn’t choose a more straight forward method because it didn’t give them the results they wanted. I guess it shows that for a selected group of states, Clinton support tended to come from the large precincts that may have taken a while to tally, but these may also be heavily urban precincts with large minority populations so its not surprising.

22

20,000+ emails say the Democrat primary campaign may not have been totally free and fair. There does exist that possibility.

23

I’ve wondered why Bernie supporters find it odd that the Democratic establishment would not be behind someone who has refused to be a Democrat until deciding to run for president. Maybe those who have devoted their lives to the party have a problem with that?

24

For example, I’m having trouble with statements like this

Based on those numbers the effect size looks somewhere around 0.25 and the number of people polled/voted is not given but I would question those t-test results too. I ran the numbers in an online calculation (without the N’s of course) and it came up as not significant at 5%. Lastly, let’s just say the numbers do say that their was only a 0.2% probability that Hillary’s actual count is not significantly different than the exit polls - how does that even come close to 1 in 77,000,000,000?

25

Not only that, but you only get delegates if the number of people choosing you is a “viable” number. How many of O"Malley’s supporters were forced to change their candidate (many to Sanders IIRC) during the early caucuses. THAT would skew the numbers too.

26

Actually, I think it was Rajneesh, and, bless me, it was a 70- or 80- ton steel beam in the (obviously fake) photo. Chinmoy doing 2200 pounds with a standing lift is impressive, and is, at least, within an order of magnitude of conceivable human limits. The scrawny and wizened Rajneesh under the many-ton beam was only pathetic. (I’ve looked and look for a copy of the photo and never found it.)

Wild exaggeration usually loses the attention of intelligent people. If a critic says, “Clinton is not competent to be President,” that’s ordinary hot air. But if he says, “She’s the worst candidate ever to run,” that’s really stupid and we know better. If he says, “Clinton should be prosecuted for distributing secure data,” that’s within the realm of sanity, but if he says, “She’s a traitor and needs to be executed,” that’s nothing but nuts.

27

Given that Clinton did measurably better among minority groups than Sanders, and that the South has a higher percentage of minority voters, this sounds dangerously close to saying Black people are stupid.

And that lack of cross-demographic appeal is one of the big things Sanders supporters don’t get: Sanders did really well among white progressives and tepidly or even badly among every other group within the Democratic Party. Sorry, you can’t win a DNC primary with only one demographic. If you want to do that, join the GOP and appeal to low-education low-income white people.

28

QFT. I was told on this board that blacks were “low information voters,” and that Clinton “doesn’t really have a base.”

“But–but everybody I know voted for Sanders!” isn’t revealing of anything, except that perhaps your social circle is too small and not nearly diverse enough.

(Actually, the next line is usually “all the blacks and Hispanics I know voted for Sanders, too!,” which doesn’t actually help.)

29

I actually think a lot of people experience political monoculture in their personal relationships now. When I was a kid, I’d say politics was more partisan, in that we had fewer independents, more of a sense of representing your family as big D or big R. But people were less politically polarized. The next door neighbors were Republicans, that was one fact about them. Political discussions weren’t part of dinner conversation. My dad and their dad were good friends despite being straight ticket voters of different parties. [This is a sketch, not my real experience.]

There’s been a lot of articles I’ve read recently, that more and more, people won’t even date someone if they have different political views. So the very frequent experience I can recall, of a husband and wife being “ticket splitters”, is now getting less common. And we aren’t just discriminating in our sexual relationships, but in our friendships as well.

I think it’s very much the case that most people simply do not seek to engage in close friendships with people who have different political beliefs, and they choose not to maintain with much vigor any friendships they have across cross-political spectrum lines.

30

Motherfuckers acting like they forgot about Kael.

Martin Hyde, Ulf the Unwashed, none of this is new. Noted film critic Pauline Kael said this, right after one of the biggest Presidential landslides in our history:

This kind of political sorting isn’t new, because politics as a proxy for cultural alignment isn’t new, and people tend to make friends with people they feel similar to. You can argue that it’s worse now than it was in the past, but 1972 was a referendum on Nixon’s Silent Majority versus the Anti-War Anti-Draft Pro-Civil Rights long-haired dope-smokers who made the evening news.

Well, these days the Silent Majority isn’t low-education white people. As per FiveThirtyEight, the modern group which most closely approximates the Nixon-era Silent Majority (white Christians without a college degree) amounts to one-third of voters. In terms of marketing demographics, that’s quite a lot. It’s more than enough to keep a few TV networks running, especially in this, the twilight of the broadcast media era, when ratings numbers which would have meant swift cancellation thirty years ago are cause for jubilation. It’s enough to keep them in a nice little bubble where they think they’re the majority. But it isn’t enough to push Trump over the top. The idea of bubbles being a purely Liberal thing is as dated as the idea that a candidate who appeals to whites alone can win the Presidency.

31

Have you actually read the letters? A lot of them are complaining that the DNC was doing too much for Sanders and should be doing more for Clinton. Sanders supporters can read that as evidence of a anti-Sanders bias. But it can equally well be seen as evidence of a pro-Sanders bias.

32

I think that “one in 77 billion chance” is a gross exaggeration. But there was definitely a split, contentious in some places, between party workers who supported Sanders and those who supported Clinton.

As I understand it, there actually was some primary election fraud in Chicago favoring Clinton. Was there enough of that around the country to rig the whole primary? Maybe not. But having any is a problem, and Sanders supporters have a right to be offended.

33

Although it doesn’t show up from the statistics presented the p=0.002 p-value comes from a paired t-test looking at the differences between votes vs exit polls, its a legitimate way to do statistics, but the effect is so small that it can easily be explained otherwise. the 77 billion quote comes from the analysis of plots ordering the votes by size of district. Since I don’t even understand the point they are trying to make with that data, I didn’t bother looking too close at how it was calculated.