United States presidential election, 2012

This page is a encyclopedia article. The page identifier is Op_en5845
Moderator:Jouni (see all)
This page is a stub. You may improve it into a full page.
Upload data {{#opasnet_base_link:Op_en5845}}

In August 13, 2012, Francois Choquette and James Johnson published a paper claiming that there are such features in the Republican primary election results that are statistically implausible to occur in an election that has not been manipulated.^[1] In Finland, the issue was first raised by the Facebook group Open Democrary Finland on October 31, 2012.

Websites describing the situation

Research plan

The reseach question is this:

Based on statistical analysis of the election data, is there evidence of fraud in the US presidential election, 2012?

Collect large enough a group of volunteers capable of a) data management, b) statistical analysis, c) wiki working, d) dissemination.
Work with the primary election data until the real presidential election data comes out on Tuesday.
Develop data management systems for voting data in Opasnet.
Develop statistical analyses (based on ^[1]) in Opasnet.
Agree on division of tasks for the presidential election.
When the presidential election data becomes available, manage and analyse the data immediately based on the task devision.
Publish results widely.

Data

Data of Presidential Election 2012, including Republican Party Primary Election.
State		Primary election	Presidential election
Alabama	AL	Precincts on county files [2]
Alaska	AK
Arizona	AZ	Precincts on county pages [3] Precincts
Arkansas	AR
California	CA		Counties
Colorado	CO
Connecticut	CT
Delaware	DE
Florida	FL
Georgia	GA
Hawaii	HI	[4] Data
Idaho	ID
Illinois	IL
Indiana	IN
Iowa	IA
Kansas	KS
Kentucky	KY	[5] [6] [7]
Louisiana	LA	[8] Counties separately [9] Counties
Maine	ME
Maryland	MD
Massachusetts	MA
Michigan	MI	Counties
Minnesota	MN
Mississippi	MS
Missouri	MO
Montana	MT
Nebraska	NE
Nevada	NV	[10]
New Hampshire	NH	Excel, precincts on county sheets Data
New Jersey	NJ
New Mexico	NM
New York	NY	[11]
North Carolina	NC
North Dakota	ND
Ohio	OH	Excel for precincts 6th March block level election data created by Mark Salling [12]	County level
Oklahoma	OK		Zip 6th March
Oregon	OR
Pennsylvania	PA		Chester county
Rhode Island	RI
South Carolina	SC
South Dakota	SD
Tennessee	TN
Texas	TX
Utah	UT
Vermont	VT
Virginia	VA
Washington	WA
West Virginia	WV	[13] Counties\|\| Precincts in county-specific files 8th May\|\|
Wisconsin	WI
Wyoming	WY

From the Coquette and Johnson paper

US Census Bureau: Census 2000 U.S. Gazetteer Files. County locations: [14]
US Census Bureau: Population, Housing Units, Area, and Density 2010. Broken?
Iowa Election Results, January 3, 2012: [15]
New Hampshire Election Results, January 10, 2012: [16]
Arizona Election Results, February 28, 2012: [17] [18]
Ohio Election Results, March 6, 2012: [19]
Oklahoma Election Results, March 6, 2012: [20]
Alabama Election Results, March 13, 2012: [21] [22]
Louisiana Election Results, March 24, 2012: [23] [24]
Wisconsin Election Results, April 3, 2012: [25]
West Virginia Election Results, May 8, 2012: [26]
Kentucky Election Results, May 22, 2012: [27]
[28] [29]

Unsuccessful searches of the claimed fraud

The claim that a major presidential candidate is in the finals because of an election fraud is a very severe one. One could assume that this would make headlines, if evidence is strong. However, nothing has been found from the major U.S. daily newspapers.

The Wall Street Journal
USA Today
ABC News
New York Times
Los Angeles Times
Washington Post
Chicago Tribune
Wikipedia talk page about the election (fraud issue raised 29th October, no activity since)

Calculations

With this code, you can reproduce parts of the original Choquette and Johnson paper.

+ Show code - Hide code

library(OpasnetUtils)
library(ggplot2)

###### Graph functions

fig <- function(data, candidate, log = TRUE) {
	out <- ggplot(data[data$Candidate == candidate , ], aes(x = Tally, y = Support)) +
		geom_point(shape=1) +    # Use hollow circles
		geom_smooth() +           # Add a loess smoothed fit curve with confidence region
		geom_smooth(method = lm) +
		opts(
			axis.text.x = theme_text(size = 20), 
			axis.text.y = theme_text(size = 20), 
			axis.title.x = theme_text(size = 20),
			axis.title.y = theme_text(size = 20, angle = 90), 
			legend.text = theme_text(size = 20),
			legend.title = theme_text(size = 20),
			title = paste("Smoothed regression of support of ", candidate, " along the size of precinct"),
			plot.title = theme_text(size=20)
		) + 
		scale_y_continuous("Candidate result in precinct, %") +
		if(log) {scale_x_log10("Vote tally in precinct") } else {scale_x_continuous("Vote tally in precinct") }

	return(out)
}

cumfig <- function(data, candidate) {
	out <- ggplot(data[data$Candidate == candidate , ], aes(x = Cumtally, y = Cumsupport)) +
		geom_point(shape = 1) +
		geom_smooth()
	return(out)
}

## OHIO ELECTION DATA NOV 2012

Fetch2(data.frame(Name = "election.ohio.2012", Key = "SSDDvuDfVDKOhlkO"))
colnames(election.ohio.2012@data)[colnames(election.ohio.2012@data) == "REGISTERED_VOTERS"] <- "Tally"

election.ohio.2012@data <- merge(
	election.ohio.2012@data, 
	as.data.frame(as.table(tapply(election.ohio.2012@data$Result, election.ohio.2012@data[c("County", "Election")], sum)))
) # Total votes for each county and election.


election.ohio.2012@data$Support <- election.ohio.2012@data$Result / election.ohio.2012@data$Freq


fig(election.ohio.2012@data, "Obama, Barack (D)") 
fig(election.ohio.2012@data, "Romney, Mitt (R)")


# OHIO PRESIDENTIAL PRIMARY ELECTION DATA

#data.oh <- opasnet.csv("c/c8/RepublicanPrimaryElection2012_8-StateCountyTotals.csv")
#head(data.oh)

# OKLAHOMA PRESIDENTIAL PRIMARY ELECTION DATA

data.ok <- opasnet.csv("6/6e/PrimaryElection2012OKresults_20120306.csv", sep = ",", header = TRUE)

data.ok <- data.ok[data.ok$Race_desc == "FOR PRESIDENT" & data.ok$Race_party == "REPUBLICAN" , ]
data.ok$Cand_desc <- data.ok$Cand_desc[ , drop = TRUE] # Drop unused levels of candidate names.

data.ok$Precinct <- as.factor(data.ok$Precinct) # Change precinct numbers to factors.

Totals <- as.data.frame(as.table(tapply(data.ok$Total_votes, data.ok["Precinct"], sum))) # Total votes for each precinct.
data.ok <- merge(data.ok, Totals)
data.ok$Support <- data.ok$Total_votes / data.ok$Freq

# Order precincts from smaller to larger and calculate candidate-specific cumulative sums.

data.ok <- data.ok[order(data.ok$Freq), ]
for(i in levels(data.ok$Cand_desc)) {
	data.ok$Cumvote[data.ok$Cand_desc == i] <- cumsum(data.ok$Total_vote[data.ok$Cand_desc == i])
	data.ok$Cumtally[data.ok$Cand_desc == i] <- cumsum(data.ok$Freq[data.ok$Cand_desc == i])
}

colnames(data.ok)[colnames(data.ok) == "Freq"] <- "Tally"
colnames(data.ok)[colnames(data.ok) == "Cand_desc"] <- "Candidate"


# Calculate cumulative support

data.ok$Cumsupport <- data.ok$Cumvote / data.ok$Cumtally

ggplot(data.ok, aes(x = Cumtally, y = Cumsupport, colour = Candidate)) +
		geom_line(size = 1.2) +
		opts(
			axis.text.x = theme_text(size = 20), 
			axis.text.y = theme_text(size = 20), 
			axis.title.x = theme_text(size = 20),
			axis.title.y = theme_text(size = 20, angle = 90), 
			legend.text = theme_text(size = 20),
			legend.title = theme_text(size = 20),
			title = "Cumulative support of candidates, Oklahoma Primary election, March 6th, 2012", 
			plot.title = theme_text(size=20)
		) + 
		scale_x_continuous("Cumulative vote tally") +
		scale_y_continuous("Candidate result, fraction", limits = c(0, 0.5))

fig(data.ok, "JON HUNTSMAN")
fig(data.ok, "MICHELE BACHMANN")
fig(data.ok, "MITT ROMNEY")
fig(data.ok, "MITT ROMNEY", log = FALSE)
fig(data.ok, "NEWT GINGRICH")
fig(data.ok, "NEWT GINGRICH", log = FALSE)
fig(data.ok, "OVER VOTES")
fig(data.ok, "RICK PERRY")
fig(data.ok, "RICK SANTORUM")
fig(data.ok, "RICK SANTORUM", log = FALSE)
fig(data.ok, "RON PAUL")
fig(data.ok, "UNDER VOTES")

----#: . Code seems to work, but gives a huge yellow errorbox. --Heta 12:53, 27 August 2013 (EEST) (type: truth; paradigms: science: comment)

Help for geom_smooth()

References

↑ ^1.0 ^1.1 ^1.2 Francois Choquette, James Johnson: Republican Primary Election 2012 Results: Amazing Statistical Anomalies. The Money Party, 2012. [1]

Related files

[statisticalanomalies-1] 1.0 ^1.1 ^1.2 Francois Choquette, James Johnson: Republican Primary Election 2012 Results: Amazing Statistical Anomalies. The Money Party, 2012. [1]

[1]