Ricochet is the best place on the internet to discuss the issues of the day, either through commenting on posts or writing your own for our active and dynamic community in a fully moderated environment. In addition, the Ricochet Audio Network offers over 50 original podcasts with new episodes released every day.
Do Republicans Lose An Abnormal Number Of Close Elections?
I decided to actually look at some data. Before doing so I set the parameters, looking at all House and Senate elections from 2006 through 2016 and defining very close elections as those in the House where the winner won by 1.5% or less of the total vote, and those in the Senate where the winner won by 1.0% or less of the total vote.
Below is the raw data. Hopefully some Ricochetti with more expertise can tell us whether there is anything statistically significant in the results.
There were 3,178 House and Senate seats up for election during the time period.
Republicans won 1,602 or 50.4% of the races while Democrats won 1,576 or 49.6%.
Very close elections were relatively rare with only 35 during the time period or 1.1% of all elections.
Of the 35 close elections, 31 were for the House and 4 in the Senate (Montana 2006, Minnesota 2008, Nevada 2012, New Hampshire 2016).
Of the close one, Democrats won 23 or 65.7%, while Republicans won 12 or 34.3%.
Democrats won 20 of 31 close House elections and 3 of 4 Senate.
One piece of data stood out when I did the analysis. 13 of the 35 close races or 37.1% took place in 2006. Only 22 occurred during 2008, 2010, 2012, 2014 and 2016.
In 2006 Republicans won 8 of the 13 close races. Since then Democrats have won 18 and Republicans only 4. From 2008 to 2016 Republicans won 51.1% of House and Senate races but only 18% of the close ones.
The 2018 races are outside the scope of the study since results are not final but this is what it looks like as of right now:
Nine House races are within 1.5% and Democrats lead in six of the 9. In the Senate we have Arizona and Florida, while in Mississippi the Republican holds a 0.9% lead over the Democrat in a race that will go to a runoff (though the R is heavily favored in the runoff).
Another interesting study would be a more qualitative analysis of the 22 close races since 2008 focused on the comparative efforts by each party to monitor and challenge during the vote counting and recounting process.
Published in General
It’s not how many people vote, it’s how many votes get counted.
These are just close races not ones where you have a recount?
Based on final results, including any recounts.
Not sure if you can do this (from your data set) but what does it look like if you just look at recounts? Also if you can, do recounts tend to flip the result or do they confirm the initial count?
I think for statistical significance you might consider a Chi square test, since you have binary out comes (D or R). This will at least tell you if the deviation is significant given the sample size. But, I don’t think you can compare R to D wins becuase the numbers are dependent (each D win is an R loss by definition) I think you would have to ask do D’s win more in close races than they win in none close races. Or do Republicans lose in closes races more than they do in not close races. That way the numbers are independent. (Hope I explained that correctly).
The one problem though even if significance is proven is that differential results don’t prove foul play in and of themselves. Because there could be other factor that account for the trend. You would need to come up with some way of distinguishing between the various possibilities.
I agree with most of your comments which is why I was careful to explain this is raw data and needs to be tested for significance. Moreover, even if statistically significant, it would not explain the cause or causes of the results.
You can compare the outcome of the close races with the overall results which is why I reported on the total number of races (3,176) verus just the close ones (35).
Without a lot more research it is difficult to tell which of the 35 races involved formal recounts.
You would think so, wouldn’t you. But I seem to remember one argument in BvG that there must be a problem because one (Jewish) precinct couldn’t have voted that way (for Bush) so the ballot must be too complex and a manual review (and correction) process was called for.
This was Stalin, right?
Thank you for this analysis. I’ve been wondering about this too.
There doesn’t have to be, although there could be, a grand conspiracy for corruption to occur on the Democrats’ side. Rush Limbaugh used to talk about the problem of the Democrats’ always thinking and preaching that they had some sort of moral supremacy over the Republicans. The result could very well be a significant amount of low-level cheating. I think the Republicans should have addressed both of these issues–voting procedures and public relations–years ago.
As would anyone who is intelectually honest. So thank you.
The whole ballot being confusing issue from 2000 though didn’t really lead to anything. It was just another example of special pleading and whining. I would like to point out that I don’t recall many Republicans buying into the disperate impact theory in other cases where it shows up.
Under the doctrine of Disparate Impact, differential results actually can prove foul play and the need for government remediation. If you don’t believe me, believe wikipedia.
Pretty much. The supposed Stalin quote is
Just to further clarify, the raw data does show that D’s win more in close races than in non-close races. They won 65.7% from 2008-16 while winning 49.6% overall. From 2010-16 they’ve won 81.1% versus 48.9% overall.
Another area for further analysis would be a sensitivity analysis. If, for instance, the parameters were changed from 1.5% margin in House races to 2%, and in Senate races from 1.0% to 1.5% do the results change significantly?
Last year the Republicans literally held the Virginia House of Delegates by a coin toss. The deciding seat was an exact tie. There were disputed ballots, but the disputes were public — as I recall, there was a picture floating around of the crucial final ballot.
What leaves a bad taste in our mouths is that Democrats always seem to gain votes towards the end of a count. Not always (the 2016 Wisconsin recount, for example — I believe Trump/Pence gained a few votes). There’s a non-suspicious reason for that, though — it makes sense that the cities would have more votes and take longer to count them. If you sit and watch vote totals coming in (which I don’t recommend) in competitive races the Republican usually starts out ahead. It tends to narrow even in races the Republican wins easily. The Republican might be ahead 10 points in the count when they call the race, and at that point hardly anyone checks three days later to see that the final tally is closer to 5 or 6 points.
This is all by design/emergent. Dems get lumped together in districts by gerrymandering (design) or self-selection (emergent order) and that leaves a bunch of districts that are split with small GOP advantage. So in a normal year, you have a few deep blue districts and many pinkish districts and GOP winning margins are small than DNC winning margins.
Statistics are tricky with humans since they all have agency (except for NPCs on the left ;)
Does it? I would think that some effort was put into seeing that precincts had roughly the same populations.
I’m afraid the sample size is just too small.
I was wondering about that.
It also matters who votes the registered name. The point of not showing ID is so reliable operatives can vote for the names who do not, or cannot show up. This is one of the reasons it is so difficult to determine fraud after the fact.
My appologies, if I was unclear. The Chi Square Test would ask if the difference in win rate between close and normal election is significant or if it is just happen stance that you have the discrepancy. Nominally there is a difference.
It’s been many years since I read Hugh Hewitt’s “If It’s Not Close, They Can’t Cheat,” but I need to dig it out again. It started me on my interest in politics by encouraging me to volunteer as an election official and start working at polling places. As soon as she was old enough, my daughter got the training and joined me. She went on to national championships in high school debate, a degree in Politics, and running for office herself. It’s been an eventful 12 years.
In any case, Hewitt’s book, and Jonah Goldberg’s “Liberal Fascism,” made a political animal out of me. I’m grateful to both gentlemen.
I’ll add to my post that even if the data is not statistically significant, the Republican Party should look at the processes around the 22 very close elections since 2008 and identify any useful lessons, per my suggestion at the end of the OP. It should look at both the processes around initial poll monitoring and counting of ballots, as well as response once controversy is triggered. Does anyone know if the GOP has any advance planning for these situations, or a SWAT team of pre-identified resources that can be mobilized quickly?
They would need to do this for each state I think, since I imagine rules differ from State to state.
Look, Republicans are always going to have some disadvantages in the years when people swarm out to seek free stuff. That’s just human nature, and if you are speaking for responsible behavior rather than Santa Claus, it costs you.
Not in the plain numbers. Look at the type of democrat who won, or who has come close to winning. Stunningly horrific candidates. Here in AZ, the likely new senator, Cinema, was a train wreck. Past recordings of her would have pushed her out of any race, but it didn’t matter.
The classic is Ocasio-Cortez, in New York. Totally empty headed. But she beat a solid incumbent democrat. She now has no money to move to DC until her salary kicks in. She had quit her bar tending job. Stunning, this is who gets elected by democrats.
With this in mind, if dems run any decent candidate, republicans have no shot. Cities have long been lost, and now add suburbs.
Ten democrats ran for the senate in states Trump won, they have won at least 8 of those seats. So the Trump factor of 2016 is gone, and is rolling the other way.
I can see no way for Trump to win in 2020. The upper mid-west was not held in this cycle, no reason to see it come back. The election was a referendum on Trump, and he lost big.
Republicans? Mobilize quickly?
That’s funny.
And realistically, it doesn’t work. Republicans have jobs. If any of them leave their government jobs to do stuff like this, they get fired. And lately, it’s the same with high-profile private sector jobs. Democrats can usually do this sort of thing on government time, if they’re slightly sneaky about it.
What makes it all possible is Republicans voting for increases in government spending each year. That creates more Democrats who can be mobilized for stuff like this.
The problem is not illustrated by a close race per se.
First, some jurisdictions are far more corrupt than others. A close race in Nebraska and a close race in Louisiana or Illinois are very different animals. Local Democratic primary elections in Atlanta or Houston could produce levels of lawbreaking that would never happen in Vermont.
Second, how many votes look legal but are illegitimate (i.e, non-citizens, non-living, non-existent persons, multiple voting persons or just Detroit or Chicago style ballot-box stuffing) My brother knew a civic minded young man in New Orleans who routinely picked up voters in a van and drove them to the precinct to vote, then to another precinct and another…
Third, how is the counting handled? Does the count match the number of registered persons who showed up or who received a mail-in ballot and is the necessary record-keeping in place to ascertain that? How are improperly completed ballots handled? How are provisional ballots for persons with registration issues handled? Are there observers from both parties on hand in all phases of handling and counting?
Fourth, when you cheat always put the burden of proof on the party that did not cheat. For example, don’t keep a bag of bogus ballots to one side and try to prove they should be counted. Add them to the mix and force the other guy to prove they should not be included. Or in the alternative, destroy or obfuscate the records that demonstrate the bogusness of the ballot pile you want to smuggle in under the rubric of avoiding “voter suppression”.
The question is not whether Democrats cheat far more often than Republicans. They do and questioning that fact is silly. The issue is the amount and degree of cheating and whether it provided the margin in any given election. A Democrat in Pennsylvania, Michigan, Illinois or Florida running for a state office starts with an immediate tactical advantage that can easily erase a 10,000 or even 30,000 vote margin in the legitimate vote count.
Under our system a GOP candidate must do better that the margin of fraud.
[Originally said on your other thread, oops]: I’d love to see this controlled by turnout (as a % of average turnout), since it’s thought that Republicans consistently vote, and Democrats vote when the season moves them.