Voting Access and 2016 Democratic Primary Results

This is a demonstration project for an experimental visualization method I’ve been working on: 2-dimensional color mapping with colorplanes. The choropleth displays the 50 U.S states (and DC) ranked by two factors: ease of access to voting and the percentage of the vote (or equivalent for caucuses) that went to Senator Bernie Sanders whose primary opponent was Hillary Clinton.

The colorplane below the map explains the scale, colors further to the right, which have increasing amounts of blue, indicate the state was ranked by RockTheVote.org as having better voting access. Colors closer to the top, which have increasing red and decreasing green values, indicate a larger share of the vote went to Bernie Sanders.

This creates four quadrants of color to describe the states: red for low voter access states that went for Bernie, green for low voter access states that went for Hillary, fuchsia for high voter access states that went for Bernie, and blue for high voter access states that went for Hillary.

Colorplanes

The purpose of the colorplane is to increase the dimensionality of a visualization by encoding two factors into a single color that can be easily interpreted by the viewer. This is accomplished by projecting the encoded values onto a plane of YUV color space with fixed luminosity.

YUV color space was originally created with the advent of color television broadcasting. To allow backwards compatibility, the luminosity (Y) component of the color, which is the only component of black and white signals, is kept separate, and all of the color information is encoded into two additional channels (U & V). The result is that, for any fixed level of Y, the YUV color space is a plane of 4 quadrants of red, fuchsia, blue, and green (clockwise) with smooth gradient transitions between. Within each quadrant, more intense colors indicate values at the extremes and duller colors indicate values closer to the middle of the range. The four quadrants of color are analogous to the 4 quadrants of a Cartesian coordinate system, and should facilitate easy understanding of the meaning of each individual color used in the visualization.

Unfortunately, differences in display monitors can impact the perception of the colors and scale. Each of the four quadrants should be equal in size and meet in the exact center of the colorplane, which is the midpoint of the two numeric scales. However, one of my displays (a MS Surface 3) must be weak in its blue pixels because the colors appear to meet about 3/4 of the way along the horizontal scale.

Voter Access

I am using scores from the Voter System Scorecard created by RockTheVote.com. This scoring system focuses on ease of access to voting and scores states in three categories: voter registration, casting a ballot, and preparation for young voters. States score higher if voting is easier to access, and the score includes factors like same-day and online registration, early voting options, and civics education for high-schoolers.

In the visualization, states in the blue, purple, and fuchsia color ranges have easier access to voting while states colored red, brown, or green have poorer access.

Election Results

Election results are scraped from Wikipedia as of June 23rd, 2016 (California’s totals are not yet final at this time). The metric displayed is the proportion of the vote (or equivalent measure for causes) received by Bernie Sanders. Sanders carried large wins in states colored red and fuchsia while Hillary Clinton was the clear winner in states colored green, teal, or blue.

Interpretation

If voter access had a clear and consistent relationship with Sanders wins, the map would be only green and fuchsia. In the converse were true, that voter access strongly benefited Clinton, the map would have been all reds and blues.

The result was instead a mix of colors, but there is a clear lack of bright blues. This indicates that Clinton did not have any big wins in states with good voter access. Since there are both intense fuchsias and reds, Sanders on the other hand was able to achieve large victories in states with both high and low voter access scores.

There are other confounding factors to consider before judging whether voter access had an impact on the election results. Many of the low voter access states that Hillary won (green) are in the south. This region, probably not by coincidence, is also home to the highest proportions of Black voters. Whether it was the voting access, the demographics, or some other factor about the south that led to Hillary’s wins is not determinable from this analysis.

View the source on GitHub

Harry Enten’s Dirty Lie – Bernie Actually Has Big Lead with True Independents

I haven’t been posting much because I decided my time is better spent phonebanking and canvassing, but this egregious abuse of numbers brought me back for a quick one. FiveThirtyEight’s Harry Enten published an article this week titled Sanders Isn’t Doing Well With True Independents with the thesis that the 23% of self-described independents who are truly swing voters (that’s about 10% of all voters) don’t favor Bernie Sanders over Hillary or Trump. This claim is a false.

Take a look at the evidence Harry provides:

In the Gallup poll, Sanders had a 35 percent favorable rating among independents who don’t lean toward either party. Clinton’s favorable rating with that group was 34 percent. Trump’s was a ridiculously low 16 percent.

One could argue that Sanders has greater potential with these true independents than Clinton: Just 63 percent of them had formed an opinion of him, according to the Gallup poll, while 83 percent had done so for Clinton.

Wow, 34% versus 35%, they’re neck-and-neck, right? Nope. Notice how Harry only provides the percentage of voters who rated each candidate favorably but hides from you number who rated each candidate unfavorably. Any time someone gives you just the favorable ratings without their counterpart, that person is trying to deceive you.

Luckily for us, Harry provided the proportions of voters who had formed a opinion. We can infer the unfavorable rates from this as the remainder of the proportion that had an opinion minus the proportion whose opinion was favorable. That’s what you see in the chart above. Hillary’s unfavorable rate is 21 points higher than Bernie’s, so her net favorability (favorable minus unfavorable) is negative 15%. The “true independents” who dislike Hillary greatly outnumber those who like her. Bernie, on the other hand, has a net positive rating of 7%, 22 points higher than Clinton.

Hillary’s net favorability is negative 15%. Bernie, on the other hand, has a net positive rating of 7%, 22 points higher than Clinton.

As for Trump, sadly Harry did not give us Trump’s decided percentage (and the source he cited doesn’t actually provide these figures), but as long as Trumps’ unfavorability rating is at least 10%, Bernie will hold the lead with our “true independents,” and that’s a pretty safe bet.

It’s disappointing to see a writer on website that touts its rationality to be torturing data to reinforce his foregone conclusions instead of forming views based on what the data really says.

Why Vox’s Numbers for Bernie Sanders’s Tax Plans Are So Wrong

Having completed a thorough analysis of the impact of Bernie Sanders’s tax plans on incomes for individuals and families across the income spectrum, I was curious to see why the numbers in this article by Dylan Matthews in Vox were so different. Reviewing his analysis, I found that there were some analytical decisions we differed on, a few tax policies that Mr. Matthews omitted or implemented inaccurately, and some significant errors in his math. The result is that Matthews’s exaggerated all of the rates in Bernie’s plans. Even if we grant his decision to pool employee and employer taxes together, which I contend is misleading for a general audience, it turns out that his infamous 77% figure for the top bracket exceeds the real rate by nearly 10 points.

The figure above includes an accurate representation of total effective rates with taxes and average family healthcare costs included. As in Matthews’s analysis, I have pooled both the employee and employer contributions. While this gives a complete picture of the impacts of Bernie’s plans in comparison with current rates,the rates can be misleading if you’re curious what these plans mean to you. Some portions of the costs shown are paid out of an employee’s paycheck, but other portions are paid by employers behind the scenes. Additionally, the net differences (savings and increases) include the pooled amounts saved or paid by both employers and employees. If you’re looking for figures that relate more directly to what you could expect to save or pay, take a look at my previous post on Bernie Sanders’s tax plans. Otherwise, continue reading below for the details of where Mr. Matthews went wrong in his analysis.

Continue reading “Why Vox’s Numbers for Bernie Sanders’s Tax Plans Are So Wrong”

More About Bernie’s Taxes: Bachelors, Businesses, & Billionaires

In the time that has passed since my original Just How Much Would Bernie Sanders Tax Me? analysis,the Sanders campaign has experienced a virtual tie in the Iowa Caucuses and a significant win in New Hampshire. This makes an accurate perspective on Senator Sanders’s proposals even more important, as the campaign draws more scrutiny due to its demonstrated ability to win and the proposals become an increasingly real possibility for the American people.

My initial analysis found that 70% of families would save money under Senator Sanders’s plans, but it focused on a narrow perspective: take home pay for a family of four in the bottom 95 percent. There have been many discussions, colorful comments, and requests regarding other family situations and other ways of assessing tax burden, and I have enhanced my R code to be able to address these requests. In this post, I will discuss the impact of Bernie Sanders’s tax plans in the following contexts:

  • Take home pay for single men and single moms
  • Take home pay for very large income levels
  • Tax burden for employers

Continue reading “More About Bernie’s Taxes: Bachelors, Businesses, & Billionaires”

Just How Much Would Bernie Sanders Tax Me?

This post has been edited to correct an overestimation of taxes under Bernie’s Social Security tax plan. A detailed log of changes is available on Github (04/21/2016)

There are a lot of misleading and false rumors being spread about democratic presidential candidate Senator Bernie Sanders’s tax plans (TODO: link to Mom’s Facebook page for examples). With the recent release of his final tax plan, Medicare for All, Bernie has now described in detail precisely how each of his programs would be paid for. With this information, I have completed a comprehensive analysis of the impact that Bernie’s plans would have in the incomes of American families. The result is that more than 70% of families would save significant amounts of money under Bernie’s plans. The greatest savings would be for a family with an annual income of $34,000 or 140% of the Federal Poverty Level. This family would have $8,933 more in their pockets each year. The median family, with an income of $66,633, would save $7,729 per year.

Total burden of federal taxes and healthcare for a family of 4 filing joint taxes
Total burden of federal taxes and healthcare for a family of 4 filing joint taxes

Continue reading “Just How Much Would Bernie Sanders Tax Me?”