Presidential Election Voter Turnout And The Outcome Of The 2016 Election

Executive Summary

Republican Turnout Was Disappointing, But Democratic Turnout Was Truly Dismal

Between 2008 and 2016, the share of eligible voters voting for the Republican nominee fell 6.7%, while the share of eligible voters voting for the Democratic nominee fell 17.7%. These figures are derived from an extended analysis of the number of eligible voters, as opposed to the total population or number of likely voters in both 2008 and 2016. That analysis also casts light on the importance of policies like felony disenfranchisement in some key battleground states that worked to the detriment of Democrats.

These numbers show a stark divide between the parties in a year where the public had an unfavorable opinion of both major party nominees throughout the election.  But, it overstates slack in Republican turnout because its main demographics are shrinking in the share of eligible voters over the last eight year, while it understates the degree to which Democrat voter turnout collapsed in 2016, because the main Democratic party demographics have been growing since 2008.

Republican Voters Are More Reliable And Loyal To Their Party Than Democratic Leaning Voters

Donald Trump won at a time when the an economic situation that was more or less neutral between the two political parties based upon historical experience, although it may have slightly favored Republicans. The overwhelming rejection of Trump by the mainstream media, by many corporate elites, and even by significant elite figures in the Republican party.

Donald Trump won, despite being an unprecedentedly bad major party nominee for the Presidency, because Republican voters have had very consistent and solid turnout over the years despite a lack of a good field organization in 2016, and because Republican voters collectively and overwhelmingly decided to put their partisan allegiances above the failings of their particular candidate this year. 

In contrast, Democratic leaning voters are notoriously fickle and failed to turn out to vote for Hillary Clinton at previous levels, despite a par for the course field organization and more professional campaign management, because a large share of Democratic leaning voters were uninspired by her candidacy and do not have such unwaivering partisan loyalty.

The Electoral College Is Rigged In Favor Of Republicans

Trump's win also highlights the fact that under the current electoral college system, a Republican can lose to a Democrat in the popular vote by a non-trivial margin (about a million votes in 2016) and yet still win the electoral college. Basically, the current system is rigged in a way that requires the Democrat to win 52%-54% of the popular vote under reasonably normal circumstances, to win the Presidency.

Trump's Run Had No Coat Tails, But Clinton's Poor Performance Hurt Down Ticket Democrats

While Donald Trump had few coat tails in down ticket voting which saw Republicans lose a modest number of seats in both the House and the Senate, the poor Democratic turnout caused by Clinton's uninspiring campaign also denied Democrats the Congressional gains that they would have secured with a strong candidate (which looked like they might materialized even on the eve of the election).

Bad Polling Led To Bad Tactical Choices

It is also, of course, notable, the polling nationwide conducted by a great many firms, almost across the board, greatly underestimated the amount of support that Donald Trump had with likely voters for reasons that are still to be determined. This reduced the urgency of Democratic field operations and Democratic leaners felt need to vote, in states that turned out to be close but were not expected to be so close, and also led Clinton to misallocate campaign resources in the manner that she would have if she had had access to more accurate polling information.

Further Research

It will be worth the effort to look at turnout data more closely on a state by state basis in battleground states to discover the most important trends driving low Democratic turnout in 2016.

First Snow Possible Tomorrow In Denver

"Denver hit 78 degrees today, breaking a record high set in 1941. But tomorrow, we're looking at highs in the 40s and snow!", according to Denver 7 news.

The earliest recorded date for the first measurable snow of the autumn in Denver ever measured was on September 3, 1961. The latest date ever recorded in Denver was on November 21, 1934. On average, Denver's first measurable snowfall of the year takes place on October 18th. The latest first measurable snowfall in Denver in the last ten years fell on November 15, 2010, a date we have already passed.

Today, it is the evening of November 16, 2016 and it is sixty-two degrees out, just five days shy of a tie of the record.

We are forecast to having freezing temperatures tomorrow and the day after in the late afternoons and evenings, after which temperatures are supposed to warm up above freezing even at night for several more days. There is a 60% of precipitation in some combination of rain and snow tomorrow in Denver, but a much lower chance of precipitation the day after when it could also get cold enough to snow.

So, we have about 50% odds of getting a little bit of short lived snow tomorrow (still 30 days later than usual and later than any year in the last decade or more) and about 50% odds of breaking the all time record for a late first snowfall in autumn in Denver.

In another 24 hours we will find out which scenario actually happens.

Siberia, Land Of Opportunity, Who Knew?

A 29-year-old U.S. citizen has been deported from Russia after illegally entering the country in search of “a better life,” Russian media outlets reported this week. 

The man, identified as Colorado-based insurance salesman Julio Prieto, is reported to have attempted to enter Russia via its southern border with Kazakhstan. However, after being denied entry because he lacked a visa, Prieto attempted to sneak over the border and was detained by border guards on Sept. 14 near a checkpoint in the town of Karasuk, east of Novosibirsk. 

The local prosecutor’s office said that Prieto told them he had been “looking for a better life,” Ria Novosti reported in October, with hopes of finding employment in Siberia. A number of reports noted that Prieto had been born in Mexico and that he spoke English with a thick accent. 

Prieto pleaded guilty to the charge of illegally entering Russia, for which he could have faced up to two years in prison. However, he was fined only 7,000 rubles ($105), and then declared exempt from that fine and given a lesser fine of 2,000 rubles ($30). 

On Monday, Prieto was deported from Russia on a flight from Novosibirsk to New York, with stops in Moscow and Dublin. The Russian government paid for the flight.

From here.

When Novosibrisk, Russia looks more like a land of opportunity than Colorado, either our economy isn't as good as it seems, or the key who thinks that is ill informed.

What Does It Take To Understand The Brain?

A provocative new think piece asks whether the experimental neuroscience program is on track by asking the question, "Could a neuroscientist understand a microprocessor?"

It finds that while our current neuroscience research agenda is broad enough, carried to its logical conclusion, to tell us a great deal, that it has serious blind spots that make it incapable of reaching key insights into the functioning of our brains with any amount of data collected in the current paradigm, no matter how much we gather. Therefore, neuroscience needs to supplement its research agenda with new approaches calculated to gain insight into qualitatively different kinds of knowledge about the brain.

There is a popular belief in neuroscience that we are primarily data limited, and that producing large, multimodal, and complex datasets will, with the help of advanced data analysis algorithms, lead to fundamental insights into the way the brain processes information. 

These datasets do not yet exist, and if they did we would have no way of evaluating whether or not the algorithmically-generated insights were sufficient or even correct. To address this, here we take a classical microprocessor as a model organism, and use our ability to perform arbitrary experiments on it to see if popular data analysis methods from neuroscience can elucidate the way it processes information. Microprocessors are among those artificial information processing systems that are both complex and that we understand at all levels, from the overall logical flow, via logical gates, to the dynamics of transistors. 

We show that the approaches reveal interesting structure in the data but do not meaningfully describe the hierarchy of information processing in the microprocessor. This suggests current analytic approaches in neuroscience may fall short of producing meaningful understanding of neural systems, regardless of the amount of data. Additionally, we argue for scientists using complex non-linear dynamical systems with known ground truth, such as the microprocessor as a validation platform for time-series and structure discovery methods. 

Eric Jons and Konrad Kording, "Could a neuroscientist understand a microprocessor?" (Pre-Print November 14, 2016). doi: http://dx.doi.org/10.1101/055624

Regional Culture Runs Deep

The differences in adolescent sexuality and family structure we see in "Red State/Blue State" comparisons in the past decade were deeply ingrained already in colonial New England, Pennsylvania, Appalachia and Virginia by the 1770s and have clear British antecedents which have faded to near irrelevance to some extent where they originated. 

By then, 10% of women in the Delaware Valley (New Jersey, Pennsylvania, Delaware and Northern Maryland), 15% of New England women, 30% of women in Virginia and 40% or more of women in Appalachia (one contemporaneous source put the figure as high as 94% in one county) were pregnant when they married. On that wedding day, the average Delaware Valley woman was 24, the average New England woman was 23 years old, the average Virginia woman was 18, and the average Appalachian woman was 19. About 33% of the Delaware Valley women were literate, as were 50% of the New England women, 25% of the Virginia women, and a smaller percentage of the Appalachian women.

From a four year old post at this blog discussing David Hackett Fischer's book Albion's Seed (1989).

A Profile Of Incarcerated Colorado Felons

This 2009 data on the characteristics of felons who are incarcerated in Colorado from a previous post at this blog, bears repeating:

Educationally, just 1% of those admitted to prison had an associates degree or more education although about 11% have some college, while 37% lacked a high school diploma with 36% being at least functionally illiterates who needed adult basic education instruction, rather than high school level GED instruction which would be too advanced for them. About two-thirds of those with either a high school diploma or GED had a GED rather than a high school diploma. So, less than a quarter of Colorado prison inmates graduated from high school in the ordinary course. In Colorado as a whole, 11% lack a high school diploma or GED, 89% of the age 25+ population has at least a high school diploma or GED, 65% have at least some college, 43% have an associates degree or higher degree, and 33% have a bachelor's degree.

About 8% had an IQ of under 81. A moderate to severe mental health problem is an issue for 30%. A moderate to severe substance abuse problem is an issue for 79%. A moderate to severe medical problem is present in 15%. Sex offenders make up 11% with another 5% suspected of having sex offense histories who are not convicted. An absence of adequate skills to get a job is a factor for 42%. Mental health needs differed considerably based on gender. A moderate to severe mental health problem was an issue for 22% of men and 55% of women. 

The DOC doesn't include crosstabs in its annual report or relate needs data to recidivism data, although some data along that line are collected in a separate report and here. Offenders with mental health issues are slightly more likely to lack of high school diploma or GED (31%-32% v. 28%), to lack job skills (94% v. 91%), to be sex offenders (22%-24% v. 18%), to have substance abuse problems (80%-83% v. 78%) and to have anger issues (40%-41% v. 39%) than other inmates. They are much more likely to have medical problems (25%-28% depending on severity v. 16%), to have IQ below 81 (about 8% v. 4%), and to have suicidality issues (about 21% to 30% depending on severity v. 9%). Only 30% of inmates without a substance abuse problem have a high school diploma and 24% have neither that nor a GED.

Some mental health data don't make much sense. Those who were classified as having mental health issues often had prior psychiatric hospitalization (18%-24% depending on severity) and out patient mental health treatment (42%-47% depending on severity), but among those not classified as having mental health issues, 5% had prior psychiatric hospitalization and 27% had prior outpatient mental health treatment, suggesting significant underdiagnosis of mental health issues by the DOC. Among those with mental health issues 23%-34% had a history of psychotropic medications, but so did 4% of those not so classified. Notably, less than 1% of inmates with mental health issues had a prior not guilty by reason of insanity case. 

The most common mental health conditions were drug addiction, depression, bipolar disorder, anxiety disorders, alcoholism, schizophrenia and psychotic disorders, dsythmic disorders, "disorders usually diagnosed in childhood" like ADHD, and "sexual and gender identity disorders" 1%. In all 34% of disciplinary violations were attributed to the 25% of inmates classified as having mental health issues in the detailed study on the issue, and these inmates were much more likely to be in solitary confinement or "close" supervision than other inmates (23%-24% v. 11%), despite generally similar offense severity. 

The overall percentage of inmates with moderate to severe needs in some category other than job skills (which almost all inmates seem to lack) is probably in excess of 90%, and once job skills are considered is probably in excess of 95%. 

The DOC also doesn't detail good time forfeitures or gang crime connections in its annual report, although it tracks both. About 7% of Colorado inmates are eligible for deportation upon release because they are not U.S. citizens. About 9% are foreign born (the same as the 9% of the general Colorado population that is foreign born foreign born), but the remainder are U.S. citizens not eligible for deportation. Colorado's inmates are 45% Anglo, 32% Hispanic, 20% African American, 3% Native American and 1% Asian. Colorado as a whole is 71% Anglo, 20% Hispanic, 4% African American, 1% Native American and 3% Asian.

The reality is that strongest risk factors for incarceration are (1) prior incarceration, (2) membership in a gang, (3) dropping out of high school without job skills, (4) substance abuse problems, (5) being male, (6) being black or Native American, and (7) being a woman with mental health issues.

Put another way, rightly or wrongly, someone who has never been previously incarcerated, graduated from high school in the ordinary course (or better yet an associate's degree or better), has job skills, does not have a substance abuse problem, is not black or Native American, and/or is a woman without a mental health problem, is highly unlikely to be incarcerated.

One Colorado Judge Not Retained

Periodically, all Colorado judges face retention elections. This year all state voters considered one State Supreme Court Justice and ten Court of Appeals judges (all eleven of which were recommended for retention). Voters considered a total of 61 District Court judges on a district by district basis of which two of the sixty-one District Court judges facing retention elections were recommended for retention, an two receive "do not retain" recommendations.  And, voters considered on a county by county basis, thirty-five County Court judges, all of whom were recommended for retention. In all, 107 judges faced retention elections this year, but only two were not recommended for retention.

The longest judicial retention election ballot was in Denver, where voters considered all 11 appellate judges and 18 district and county court judges, all 29 of whom were recommended for retention. El Paso County voters considered 9 district and county court judges, Jefferson County voters considered 8, Arapahoe and Adams County voters considered 7, Douglas, Elbert, Lincoln, Teller and Broomfield county voters considered 6, Larimer County voters considered 5, Clear Creek, Eagle, Lake, Summit, Mesa and Weld County voters considered 4, and voters in each of the other 46 counties in Colorado considered three or fewer District and County Court judges.

Just two of the judges facing retention this year were given "do not retain" recommendations and one was retained anyway. Just three of the last ten judges recommended to not retain were removed (receiving an average 54 percent of the vote). According to 9News:

Judge Jill-Ellyn Straus will lose her job as judge in the 17th judicial district (Adams & Broomfield counties) when her current term on the bench ends in January. 52 percent of voters decided to give Straus the boot. Her review flagged concerns about her demeanor in court and gave her poor marks on fairness and communication. More than 177,000 cast votes on the question of retaining Straus. 

The other Colorado judge to earn a “do not retain” recommendation in the state’s official voter guide kept his job. Fifty-seven percent of voters in Southeast Colorado chose to keep 16th district (Bent, Crowley, and Otero counties) judge Michael Schiferl despite a review that found he appeared to chummy with the people in his courtroom along with poor legal writing and a lack of willingness to accept constructive criticism. Fewer than 8,000 voters cast votes in Schiferl’s more rural district. 

The Case For Reform Of Judicial Retention Elections

Routine judicial retention election races clutter the ballot and reduce scrutiny of judges who really need it.

I would favor a change to the system in which judicial retention is only placed on the ballot in cases (1) where judges receive a "no recommendation" or "do not retain" recommendation, (2) a supervising appellate court asks that a judge be placed on the ballot, (3) there is a petition to put a particular judge on the ballot, or (4) a majority of relevant politicians (a majority of either house of the state legislature for appellate judges, a majority of county commissioners in the judge's district for most judges, and a majority of city council for Denver District Court, Denver Probate Court, Denver Juvenile Court, or Denver County Court judges) request that a particular judge be placed on the ballot. 

No politician who is entitled to appoint judges would have a say on eligibility for retention and the ultimate decision to retain or not would remain in the hand of voters.

This would dramatically shorten the ballot with virtually no change in outcomes, while focusing public attention on the judges whose performance actually does need to be reviewed. There are extremely few (if any) cases where a judge recommended for retention has not been retained, and no appellate judge in Colorado has ever not been retained.

Deep Genetic Links Exist Between Common Severe Psychiatric Conditions

A several notable genetic markers are common across schizophrenia, major depressive disorder and bipolar disorder according to a new meta-analysis.

The search for biomarkers has been one of the leading endeavours in biological psychiatry; nevertheless, in spite of hundreds of publications, hardly any marker has proved useful in clinical practice. To study how biomarker research has progressed over the years, we performed a systematic review of the literature to evaluate (a) the most studied peripheral molecular markers in major psychiatric disorders, (b) the main experimental design features of studies in which they are proposed as biomarkers and (c) whether their patterns of variation are similar across disorders. 

An automated search revealed that, out of the six molecules most commonly present as keywords in articles studying plasmatic markers of schizophrenia, major depressive disorder or bipolar disorder, five (BDNF, TNF-alpha, IL-6, C-reactive protein and cortisol) were the same across the three diagnoses. An analysis of the literature on these molecules showed that, whilst 66% of original articles compared their levels between patients and controls, only 35% were longitudinal studies, and only 10% presented an evaluation of diagnostic efficacy, a pattern that has not changed significantly over two decades. 

Interestingly, these molecules varied similarly across the three disorders, suggesting them to be nonspecific systemic consequences of psychiatric illness rather than diagnostic markers. On the basis of this, we discuss how research fragmentation between diagnoses and publication practices rewarding positive findings may be directing the biomarker literature to nonspecific targets, and what steps could be taken to increase clinical translation in the field.

Jairo V. Pinto, et al.,  "Transdiagnostic aspects of peripheral biomarkers in major psychiatric disorders: a systematic review" (pre-print November 8, 2016) doi: http://dx.doi.org/10.1101/086124

Ethnicity and Genetic Medicine

Ethnicity is relevant for interpreting genetic data on the context of genetically informed medical treatment of conditions such as cancer. And, in this context, ethnicity is a meaningful enough concept that it can be determined with 99%+ precision from 1000 Genomes data. 

However, this probably overstates the precision of the information because the 1000 Genomes data involves people with very well established distinct identities and has instances where people with more similar ethnicities, or who are admixed, must be distinguished from each other. The experience of commercial genetic ancestry research trained on the same data suggests that in more realistic data sets would have less precision unless the ethnic categories they are trying to identify are very coarse. 

Whole exome sequencing (WES) is widely utilized both in translational cancer genomics studies and in the setting of precision medicine. Stratification of individual's ethnicity is fundamental for the correct interpretation of personal genomic variation impact. We implemented EthSEQ to provide reliable and rapid ethnicity annotation from whole exome sequencing individual's data and validated it on 1,000 Genome Project and TCGA data demonstrating high precision (>99%). EthSEQ can be integrated into any WES based processing pipeline and exploits multi-core capabilities. Source code, manual and other data is available at http://demichelislab.unitn.it/EthSEQ.

Alessandro Romanel, Tuo Zhang, Olivier Elemento, Francesca Demichelis, "EthSEQ: ethnicity annotation from whole exome sequencing data" (pre-print published November 10, 2016). doi: http://dx.doi.org/10.1101/085837

Copyright v. Biology

When do copyright laws prohibit the use of third-party images without express permission of the author, in furtherance of biodiversity and taxonomy research?

A new paper at bioXriv explores the issue. The conclusion they reach is largely right, but probably for the wrong reasons.

In my opinion as a lawyer who sometimes deals with copyright law, the article takes a more expansive view of the scope of what is copyrightable than the law supports. The standardized presentation and "lack of creativity" defense is quite narrow.  For example, even photos taken on an automated basis or by the monkey subject of the picture have been held to be protected by copyright. The "lack of creativity" exception tends to apply to documents like directories and indexes.

There is a fair argument that there is an implied license to publish an image in a scholarly taxonomic publication granted to other persons engaged in taxonomy research that can be inferred from the customs and practice of the discipline. The analogy in area of real property law would be the implied license granted to postal workers, delivery persons, solicitors and other people who would like to talk to you to enter onto your front walk to your front door to leave a package or knock and request entry, even in the absence of an express grant of permission to do so.

Similarly, the scientific research purpose of the investigators could overcome this problem with a fair use defense which is read broadly in cases of legitimate scientific research, and in which similar arguments to those for an implied license could be considered. The fact that a journal's economic profit is not a major motivation for publishing scholarly work, and that citation tends to increase rather than decrease the value of scholarly work could also enter into consideration.

The abstract and citation to the article are as follows:

Taxonomy is the discipline responsible for charting the world's organismic diversity, understanding ancestor/descendant relationships, and organizing all species according to a unified taxonomic classification system. Taxonomists document the attributes (characters) of organisms, with emphasis on those can be used to distinguish species from each other. Character information is compiled in the scientific literature as text, tables, and images. The information is presented according to conventions that vary among taxonomic domains; such conventions facilitate comparison among similar species, even when descriptions are published by different authors. 

There is considerable uncertainty within the taxonomic community as to how to re-use images that were included in taxonomic publications, especially in regard to whether copyright applies. This article deals with the principles and application of copyright law, database protection, and protection against unfair competition, as applied to images. 

We conclude that copyright does not apply to most images in taxonomic literature because they are presented in a standardized way and lack the creativity that is required to qualify as 'copyrightable works'. There are exceptions, such as wildlife photographs, drawings and artwork produced in a distinctive individual form and intended for other than comparative purposes (such as visual art). 

Further exceptions may apply to collections of images that qualify as a database in the sense of European database protection law. In a few European countries, there is legal protection for photographs that do not qualify as works in the usual sense of copyright. 

It follows that most images found in taxonomic literature can be re-used for research or many other purposes without seeking permission, regardless of any copyright declaration. In observance of ethical and scholarly standards, re-users are expected to cite the author and original source of any image that they use.

Willi Egloff, et al., "Copyright and the Use of Images as Biodiversity Data" (pre-print posted November 11, 2016). doi: http://dx.doi.org/10.1101/087015

Consolation Prizes - State Ballot Issues

The one consolation prize from the 2016 election is that a number of good ballot measures passed at the state level.

Here are a few of them that are good news, omitting many financial or procedural measures and all measures that failed:

1. Minimum Wage Increases: AZ, CO, ME,  WA

2. Pro-Recreational Marijuana: CA, ME, MA, NV

3. Pro-Medical Marijuana: AR, FL, MT, ND

4. Stricter Gun Laws: CA, ME, NV, WA

5. Legalize Assisted Suicide: CO

6. Tobacco Tax Increases: CA

7. Ranked choice voting: ME

8. Downgrades Some Drug Offenses To Misdemeanors: OK

9. Beer Sales Allowed In Grocery Stores: OK

What Are The Consequences Of The Election?

As a result of the 2016 election, we will have a Republican President for the next four years who is the least competent man to hold that office of all time, a Republican controlled U.S. Senate for at least two years, a Republican controlled U.S. House of Representatives for at least two years, and a conservative controlled U.S. Supreme Court for the foreseeable future.

(Incidentally, this continues the long term trend of realignment as the Northeast has fewer Republicans in Congress, while "A third of all House Democrats now hail from three states (CA, NY, MA). California alone accounts for 20% of the House Democratic caucus.")

Republicans now control 33 Governorships up from 31 before the election, and more state legislative chamber than the 79 out of 99 that they already controlled.

This is an unmitigated disaster from a policy perspective. 

Somehow or other, Obamacare will be repealed and many millions of Americans will lose their health insurance. There will be no improvements in access to higher education.

Somehow or other there will be reckless changes to the tax code that reduce federal revenues without offsetting spending cuts driving up the deficit and favoring the rich. 

Our economy will falter as public investments decline and sensible regulations are not enacted or are rolled back. Our environment will suffer. Science will be suppressed.

All progress on social issues will go down the drain. Civil liberties will be trampled upon.

Any effort to make our elections more fair or to restrain corruption will be rolled back.

We will become isolationist and xenophobic on immigration, refugees and trade.

Our nation will become an international diplomatic pariah. 

We may very well commence World War III, recklessly, in an expressly anti-Muslim war.

It may actually be good for me personally, at least from a business development perspective, as I will have to help clients adapt to and plan around a whole range of rapidly changing federal policies. But, it will not be good for America.

Why Did Trump Win?

Twenty-Four Theories

CNN has a nice article that lays out 24 theories. These boils down to three main themes, however:

One: Democrats backed Clinton despite Sanders being a stronger candidate, and then Clinton couldn't unite and mobilize the party. Many strong Democratic candidates didn't run.

3. Because of low voter turnout

10. Because the Democratic Party establishment didn't push Bernie Sanders

12. Not because of millennials

13. Because of Gary Johnson and Jill Stein

20. Because Democrats focused more on turning out supporters than growing the base

21. Because the Democratic National Committee selected the less competitive candidate

24. Not because of Comey

Two: Voters were deceived by lies, blinded by fluff and inflamed by a demagogue.

1. He won because of Facebook and its inability or unwillingness to crack down on fake news

2. Because of social media, generally

4. Because celebrity outlasts substance

7. Because of Russia after all?

9. Because rural Midwesterners don't get out of the house enough

23. Because of Comey

Three: Democrats ignored the white working class sense that the media and system are corrupt and they don't matter and Democrats didn't consider their racist concerns legitimate

5. Because of white women

6. Because of white male resentment

8. Because the left and coastal elites shamed Trump supporters

11. Because Reagan Democrats surged in Michigan and Midwest

14. Because political correctness set off a nasty backlash

15. Because he simply listened to the American people

16. Because college educated Americans are out of touch

17. Because Americans are biased -- but not against any race, ethnicity or gender

18. Because voters believed the system was corrupt

19. Because he remembered 'forgotten men, women' of America

22. Not because of racism

Too Many Democrats Stayed Home

Turnout (voting-eligible population):

2008: 62.2%
2012: 58.6%
2016: 55.6%

Steve Greene highlights much lower turnout by supporters of Hillary Clinton than by supporters of Barack Obama, than in either 2008 or 2012, while Republican turnout has been more or less static. Clinton won the popular vote, but by a margin much smaller than Obama in 2008 or 2012.

Note that Hillary Clinton didn't need to perform nearly as well as she was polling prior to the election to win. All she needed to do was perform 1.1 percentage points better in Wisconsin, 0.4 percentage points better in Michigan, and 1.3 percentage points better in Pennsylvania. That is all she needed to win those states, even though that would still have been far worse than she polled in those states.

Sanders won in the primary in Minnesota, Wisconsin, Michigan, New Hampshire and Maine, which are all relatively close states where Clinton underperformed in the general election, although Clinton did win Pennsylvania in the primary against Sanders.

But, it could also have something to do with overconfidence inspired by inaccurate polls.

Other Reasons

* The Electoral College has a Republican bias. 

Because the Democratic party is more urban and the Republican party is more rural, a popular vote tie means an electoral vote win for Republicans. Democrats need to win by at least two or three percentage points in the popular vote in order to win the Electoral College. Hillary Clinton won the popular vote, but not by a big enough margin to defeat Trump.

* Republicans are more partisan than Democrats

This was a race with the two least popular major party nominees in history. But, Republicans didn't care and Democrats and unaffiliated voters did.

Ultimately, despite having the worse major party nominee in Republican party history, who was morally flawed, sexist, a serial rapist, a racist, a fraudster, and a man who has made many horrible, horrible business decisions, grass roots Republicans ultimately decided to overlook those flaws pointed out in the primary, because they want their party to win and support the "Alpha Male".

Democrats like a thin elite of Republicans, are choosier and didn't get as strongly behind their own flawed candidate who was not an "Alpha Female", who was not an inspiring orator or visionary, and who was unable to dispel criticisms from Democrats that she was a centrist corporate pawn and from Republicans that she was corrupt. A lack of enthusiasm from Democrats and unaffiliated voters for Clinton translated into lower turnout by supporters because they are pickier and hold their leaders to higher standards rather than blinding getting behind the cause.

* Clinton failed to inspire.

Clinton came up short in free media because she didn't give them the inspirational or at least noteworthy soundbites that make news. Without them, you don't get the coverage and hype that you need to penetrate to consciousness of the electorate.

The most inspiring slogan of her entire campaign, "a taco truck on every corner" was devised by a Republican critic and adopted by her supporters because she had nothing better to offer.

You can't win hearts and minds with twelve point plans and policy nuances. You need to communicate a vision emotionally and delegate the details to others. Without a short defining message to sum up her campaign in a few words, she let others define her in a negative light. Without opening up in her speeches, she couldn't communicate her heart and character to the American people. Presidents need to be hedgehogs and not foxes.

You can call those expectations male bias, but inspiring speeches are a Presidential candidate's stock in trade and a central part of the process. Without them, you can't win.

If you don't reassure the public with a constantly repeated message, they assume the worst and dwell on the negatives. 

It wasn't that the public has deeply embraced the Republican vision for America. While Clinton's campaign did not deliver the coattails that had been expected, the Republican majority in the U.S. Senate narrowed from 54-46 to 52-48, and Democrats gains six more seats in the U.S. House at the expense of Republicans. Few incoming Presidents lose support in both houses of Congress in the year they are elected. Trump takes office after failing to win the popular vote and winning the Electoral College by a razor thin margin in three critical states. 

But, Clinton's campaign was not rousing enough to surmount Trump's thin support, and failed to realize the coattails she could have commanded, and indeed was expected by polls to command.

* The media and political elites took too long to respond to big lies; they weren't effective

Only one major newspaper in the entire country, in Las Vegas, endorsed Trump, that paper had a personal connection and financial interest in Trump, and that endorsement wasn't enough to prevent Trump from losing in Nevada. Indeed, almost every state in which Trump has significant financial investments rejected him. His popularity fell following every general election debate.

The utter failure of this media consensus to have any meaningful impact on the outcome of the election is a profound statement about its irrelevance in modern Presidential politics.

Trump lied all day long, every day of his campaign, and the media and political elites didn't manage to start taking him seriously to task for this until the final month or so of the campaign after most people had already made up their minds. Somehow they could never made the fact that Trump was a liar or that his character, in general, was profoundly flawed stick, in large part because the political right has spent decades convincing itself that the mainstream media is biased and can't be trusted.

* Bad polling data led to bad tactical decisions

Clinton mounted a campaign tactically based on the assumption that the polls were accurate and that she had solid support in Pennsylvania and the Midwest, when in fact those states were far closer and needed far more attention. She made these decisions because the polling profession across the board failed to accurately capture the strength of Trump's support. 

What Did The Polls Get Wrong?

Why was the 2016 election win of Donald Trump so surprising?

Mostly because some key polls were considerably off the mark.

The states with the narrowest pre-election polling, followed by the actual results, follow, with errors of more than 4 percentage points highlighted.

                                  Predicted                   Actual            Underestimate of Trump
Washington                 D+13.2                   D+17.6                               (4.4)

Connecticut                 D+12.9                   D+12.2                              (0.7)
Illinois                         D+12.8                   D+16.0                              (3.2)
Delaware                     D+12.7                   D+11.5                               1.2

New Jersey                  D+11.7                   D+12.8                              (1.1)
Oregon                         D+9.2                     D+11.0                             (1.8)

Maine - statewide        D+7.8                     D+2.7                                 5.1
New Mexico                D+5.9                     D+8.3                                (2.4)
Minnesota                    D+5.7                     D+1.4                                 4.3
Virginia                        D+5.5                     D+4.9                                 0.6

Wisconsin                    D+5.3                      R+1                                    6.3
Michigan                     D+4.3                      R+0.3 (not final)                 4.6
Colorado                      D+4.0                      D +2.1                                1.9
Pennsylvania                D+3.7                     R+1.2                                  4.9
New Hampshire           D+3.6                      D +0.2                                3.4
Nevada                         D+1.2                      D+2.4                                (1.2)
Maine 2nd District       D+0.8                      NA                                     >0.8
Florida                         D+0.6                      R+1.3                                 1.9              
North Carolina             D+0.6                     R+3.8                                  4.4

Ohio                             R+1.8                     R+8.6                                  6.8
Nebraska 2nd District  R+2.3                      NA                                       NA
Arizona                        R+2.3                     R+4.3                                  2.0
Iowa                             R+2.7                     R+9.6                                  6.9

Georgia                        R+4.2                     R+5.7                                  1.5
South Carolina             R+7.2                     R+14.1                                6.9
Alaska                          R+7.8                     R+15.2                                7.4

Texas                            R+8.8                     R+9.2                                  0.4
Missouri                       R+10.0                   R+19.1                                9.1
Utah                             R+11.2                    R+19.0                               7.8
Indiana                         R+11.5                    R+19.3                               7.8

Tennessee                    R+12.4                    R+26.2                              13.8

Kansas                         R+13.0                    R+21                                   8.0
Mississippi                  R+13.1                    R+18.5                                5.4

It was not particularly notable that Trump won North Carolina, Florida and Maine's 2nd District, all of which had razor thin margins in favor of Clinton. But, Trump's performance was widely underestimated by large margins in a great many states and was overestimated in just a few states, most of which were in the western United States.

FiveThirtyEight looked at what the polls got wrong. The trends that it identified in the errors were as follows:

While the errors were nationwide, they were spread unevenly. The more whites without college degrees were in a state, the more Trump outperformed his FiveThirtyEight polls-only adjusted polling average,1suggesting the polls underestimated his support with that group. And the bigger the lead we forecast for Trump, the more he outperformed his polls.2In the average state won by Trump, the polls missed by an average of 7.4 percentage points (in either direction); in Clinton states, they missed by an average of 3.7 points. It’s typical for polls to miss in states that aren’t close, though. The most important concentration of polling errors was regional: Polls understated Trump’s margin by 4 points or more in a group of Midwestern states that he was expected to mostly lose but mostly won: Iowa, Ohio, Pennsylvania, Michigan, Wisconsin and Minnesota.

Some of the main explanations for this include the following:

James Lee of Susquehanna Polling & Research Inc. said his firm combined live-interview and automated-dialer calls, and Trump did better when voters were sharing their voting intention with a recorded voice rather than a live one. 

Women who voted for Trump might have been especially reluctant to tell pollsters, said David Paleologos of Suffolk University. The USC Dornsife/Los Angeles Times poll corroborated that: “Women who said they backed Trump were particularly less likely to say they would be comfortable talking to a pollster about their vote.” 

Gourevitch offered a theory for why polls underestimated Trump support: “that some percentage of the Trump vote is distrustful of institutions and distrustful of poll calls.” 

Pollsters also cited lower-than-expected turnout, particularly in the Midwest. “Democrats had a turnout problem,” Gourevitch said, and therefore so did pollsters. “The turnout models appear to have been badly off in many states,” said Matt Towery of Opinion Savvy. 

It also looks as if Trump pulled late support from many Republican voters who had been undecided or were supporting a third-party candidate. Libertarian candidate Gary Johnson’s recent decline coincided with Trump’s gains in the polls.

Well, That Sucks.

Unless you are a far right Republican, this year's election was decidedly disappointing, to vastly understate the situation.

So, President Trump. Either 52 or 53 Republican U.S. Senators. Republicans still control the U.S. House. Pretty much a worst case scenario election. The Denver Post headline: "World gasps in collective disbelief following Donald Trump’s election" pretty much captures it. So, Republicans will control Congress and the Presidency and will tip the balance of power in the U.S. Supreme Court.

Almost nobody (except the author of the comic strip Dilbert) predicted this would happen. Trump's strength in the Rust Belt was greatly underestimated.

The stock market plunged in reaction and the Mexican Peso is also down.

Colorado, as expected, sent its electoral votes to Clinton and re-elected its Democratic U.S. Senator, but didn't flip the 6th Congressional District seat held by Republican Mike Coffman who easily won re-election and Republican held onto control of the state senate. Democrat Beth McCann is Denver's next District Attorney. Many ballot issues were considered in Colorado:

Winning were:

Referendum T (end slavery and involuntary servitude, even as a criminal punishment)
Denver 3A and 3B (Denver Public Schools funding)
Denver Metro Area 4B (Scientific and Cultural Facilities District funding)
Denver 300 (allow private marijuana consumption clubs)
Amendment 70 (minimum wage)
Amendment 71 (making new constitutional amendments harder to pass)
Prop 106 (right to die - drug assistance)
Prop 107 (open Presidential primaries)
Prop 108 (open primaries)

Losing were:

Referendum T (end slavery and involuntary servitude, even as a criminal punishment)
Referendum U (property tax exemptions)
Amendment 69 (universal health care)
Amendment 72 (increase tobacco taxes)
Pueblo 200 and 300 (end retail marijuana in Pueblo)

The defeat of referendum T despite an all American tag line, almost no money spent to oppose it, the lack of any organized campaign against it, and bipartisan legislative support in the state house and state senate, is pretty shocking.

UPDATED WITH RESPECT TO REFERENDUM T on November 10, 2016 in response to the comments.

A Nation Divided

One of the most striking things about American politics is how deep its regional political leanings, which in broad strokes have been pretty stable since the time of the Revolutionary War if not earlier, and have been static on a county by county basis since at least 1876.

Excluding the District of Columbia and single Congressional Districts, there is a 64.3 percentage point spread in Presidential preferences and an 84.9 percentage point spread in partisan U.S. Senate preferences. The extent to which Red State and Blue State America are different countries politically is very real.

This is clearly reflected in the final state by state predictions from 538:

Final FiveThirtyEight Forecast

Who’s ahead in each state and by how much

"Our win probabilities come from simulating the election 10,000 times, which produces a distribution of possible outcomes for each state. Here are the expected margins of victory [in percentage points]"

Expected margin of victory" -26 states D and 25 states R

D.C.                                  D+71.3
Vermont                            D+27.8
Maryland                          D+25.6
Hawaii                              D+24.9
Massachusetts                  D+23.3
California                         D+22.8
New York                         D+19.2
Rhode Island                    D+15.1
Maine 1st District            D+14.2
Washington                      D+13.2
Connecticut                      D+12.9
Illinois                              D+12.8
Delaware                          D+12.7
New Jersey                       D+11.7
Oregon                              D+9.2
Maine - statewide             D+7.8
New Mexico                     D+5.9
Minnesota                         D+5.7
Virginia                             D+5.5
Wisconsin                         D+5.3
Michigan                           D+4.3
Colorado                           D+4.0
Pennsylvania                     D+3.7
New Hampshire                D+3.6
Nevada                              D+1.2
Maine 2nd District            D+0.8
Florida                               D+0.6
North Carolina                  D+0.6

Ohio                                  R+1.8
Nebraska 2nd District       R+2.3
Arizona                             R+2.3
Iowa                                  R+2.7
Georgia                             R+4.2
South Carolina                  R+7.2
Alaska                              R+7.8
Texas                                 R+8.8
Missouri                            R+10.0
Utah                                  R+11.2
Indiana                              R+11.5
Tennessee                          R+12.4
Kansas                               R+13.0
Mississippi                        R+13.1
Montana                            R+14.9
Nebraska 1st District         R+15.4
South Dakota                     R+15.6
Louisiana                           R+16.3
Nebraska - statewide         R+18.1
Kentucky                           R+18.6
Idaho                                 R+20.1
Arkansas                           R+20.8
Alabama                            R+22.3
North Dakota                    R+22.7
Oklahoma                          R+26.3
West Virginia                    R+26.4
Wyoming                           R+35.1
Nebraska 3rd District        R+36.5

"Who’s ahead in each Senate race

Our win probabilities come from simulating the election 10,000 times, which produces a distribution of possible outcomes for each race.

Incumbent party Expected margin of victory"

New York            D                D+41.1
Hawaii                 D                D+41.0
Vermont               D                D+40.7
Oregon                 D                D+27.9
Maryland              D                D+27.5
Connecticut          D                D+25.5
Washington          D                 D+16.7
Illinois                  R                 D+12.3
Colorado              D                 D+8.3
Wisconsin            R                 D+3.9
Pennsylvania        R                 D+1.2
Nevada                 D                 D+1.2
New Hampshire   R                 D+0.2

Missouri               R                  R+0.7
North Carolina     R                  R+2.2
Indiana                 R                  R+2.5
Florida                  R                 R+5.5
Kentucky              R                  R+9.7
Arizona                 R                 R+10.4
Georgia                 R                 R+12.5
Ohio                      R                 R+14.8
Arkansas               R                 R+17.7
Iowa                      R                 R+18.5
Louisiana              R                 R+23.8
South Carolina      R                 R+24.4
Alaska                   R                 R+26.4
Idaho                     R                 R+29.2
Kansas                   R                 R+30.4
Alabama                R                 R+30.7
South Dakota         R                 R+32.1
Utah                       R                 R+36.5
Oklahoma              R                 R+37.4
North Dakota         R                 R+44.8

The balance of power in the next Senate

In each simulation of the Senate elections, we forecast the winner of all 34 races and note the resulting number of seats that would be held by the parties. That gives us a distribution of possible outcomes.

R  D  Change Likelihood

42 58 +12 <0.1%

43 57 +11 0.2%

44 56 +10 0.5%

45 55 +9 1.2%

46 54 +8 3.5%

47 53 +7 7.4%

48 52 +6 10.9%

49 51 +5 14.9%

50 50 +4 16.6%

51 49 +3 15.9%

52 48 +2 13.9%

53 47 +1 9.6%

54 46 —— 4.2%

55 45 -1 1.0%

56 44 -2 0.2%

In the case of 538's Senate predictions it is worth recalling that 538 systemically overweights rare possibilities in the tails of a probability distribution, so the real range of possibility is much more like 55 to 46 Democrats with a midrange value closer than indicated, but their methods are not biased, so the average result should be about the same either way (which is a strong likelihood of the modal 50-50 split with a Vice President holding the swing vote).

Early Voting

Colorado had about 2.2 million early votes with Democratic early votes lagging Republican early votes by about 8,000 votes, far less than the 35,000 vote early voting lead that Republicans had in 2012 when Colorado ultimately cast its electoral votes for President Obama.

Lots And Lots Of Early Votes

Early voting patterns don’t tell us who will win the election, but they certainly reveal that Americans in battleground states want to have their voices heard. We now have tallies from several that show record-breaking early voting turnout, particularly among Latinos in some states. 

In Florida, 2,636,783 people voted by mail before Election Day, while 3,874,929 voted in person during early voting. That’s more than the total number of Floridians who voted in the 2000 election, period. The number of Latino early voters in the state doubled in number since the last election. 

In Nevada, 41 percent of in-person early voters were registered Democrats and 35 percent were Republicans, which does not mean everyone voted for their party of registration, of course. FiveThirtyEight’s Harry Enten wrote about the possibility that heavy early voting may have swung the state for Clinton. 

In North Carolina, the 3.1 million early absentee votes include 42 percent registered Democrats and 32 percent Republicans, which is a narrower advantage for Democrats than they had in 2012. 

Election laws vary by state, of course, but early voting is proving increasingly popular, with a record 46 million Americans voting in advance of Election Day this year.

From 538.

I think it is likely that early voting will be the biggest source of disparities between actual results and the results predicted by polls this year. Generally speaking, early voting has favored Democrats in swing states this year compared to prior years.

Almost Showtime

Election forecasts for the Presidency and the U.S. Senate have been remarkably volatile, although Hillary Clinton has still been the favorite or at closest tied with Donald Trump for most of the race. The Presidential race has been gradually tightening over the last couple of weeks, and the U.S. Senate race has dramatically tightened over the last week, according to polling.

It isn't at all clear how much of this is bad modeling and how much of this is reality. Much of the gain for Trump appears to come from reluctant, eleventh hour decisions to cease to be an undecided voter or to back the Libertarian candidate Gary Johnson. 

But, the tightening of the polls doesn't seem to comport with unprecedented early voting turnout by Democrats and may be due in part from a failure of prediction models to adequately reflect the impact of early voting and to weight mere "likely voters" who have not yet voted and tend to vote Republican properly relative to the votes of people who have already cast voters which seems to be weighted towards Clinton. For example, perhaps people who have already voted are less likely to respond to polling or are weighted equally when already voted has a 100% chance of impacting the final result, while a likely voter may only have a 90% chance of voting.

My intuition is that Clinton and Democratic candidates for U.S. Senate will over perform late polling across the board, but that could be simply cognitive dissonance at work. I find it hard to believe that Comey's disclosures at the FBI (which in the end amounted to nothing) rocked the vote very much.

In a day and a half, we will learn which predictions were on target and which were far afield.

One In Eight Cases Of Cancer Caused By Eight Kinds Of Viruses

Most cases of cancer have nothing to do with viruses. But, the role of viruses in cancer is much greater than had been widely believed not very long ago. And, quite a few environmental causes (most famously smoking or crewing tobacco) are know to be significant contributors to cancer rates. 

The US Department of Health and Human Services released its 14th Report on Carcinogens today, including seven "newly reviewed" substances, bringing its total number of known human carcinogens to 248. 

Five viruses have been added to the list:

Human immunodeficiency virus type 1 (HIV-1)

Human T-cell lymphotropic virus type 1 (HTLV-1)

Epstein-Barr virus (EBV)

Kaposi sarcoma-associated herpesvirus (KSHV)

Merkel cell polyomavirus (MCV) 

These viruses have been linked to more than 20 kinds of cancer, according to the report, including non-melanoma skin cancer, eye cancer, lung cancer, stomach cancer and multiple types of lymphoma. 

"Given that approximately 12% of human cancers worldwide are attributed to viruses, and there are no vaccines currently available for these five viruses, prevention strategies to reduce the infections that can lead to cancer are even more critical," said Linda Birnbaum, director of the National Institute of Environmental Health Sciences and the National Toxicology Program, in a statement.

From CNN.

The other viruses known to cause cancer are: Hepatitis B, Hepatitis C, and Human Papillomaviruses: Some Genital-Mucosal Types. There are vaccinations available against most strains of these viruses.

There are 240 non-viral substances that are known to cause cancer or strongly linked to cancer.

There are also many known genetic predictors of cancer risk, most famously, the genes that make one vulnerable to breast cancer.

On the other hand, a significant share of all cancers are seemingly random.

The Case For More Political Polling

In the United States in this day and age, there is exhaustive national polling in Presidential election races (including Presidential primary races) and is a significant amount of polling done in top of the ticket statewide office races (state level Presidential election polls, U.S. Senate race poll, and Governor's race polls). There is also frequency at least some statewide polling on at least the most high profile and controversial ballot issues.

But, there is generally little or no polling done on ballot issues that are not statewide, on down ticket statewide candidate races (e.g. state attorneys-general, secretaries of state, state treasurers, and various state board memberships).

There is surprisingly little polling done in races for the U.S. House of Representatives, and there is basically no polling done in races for state legislative seats and for local government offices, or even for regional government offices (e.g. in Colorado, seats on the Board of the Regional Transportation District and District Attorney elections).

A quite modest investment in additional polling would greatly improve the amount of publicly available information about the probable political consequences of upcoming elections.

The amount of information available from additional polling also wouldn't have to be very expensive to produce a lot of valuable information.

For example, an issue in the 2016 election that matters a great deal to pretty much everyone in Colorado is which party will control the state house and state senate in the Colorado General Assembly after the election.

It wouldn't be too hard to get good information on this question from polling. Half of the 35 seats in Colorado's state senate are not before the voters in 2016 since this body has four year terms with half of the seats before the voters every two years.  Voters cast ballots in all 65 seats of Colorado's state house every two years.  

But, while there are 82-83 seats in play in the Colorado General Assembly every two years, previous election results (particularly in cases where there is an incumbent running), the fact that not every seat has a candidate from both major political parties, voter registration data in each state legislative district in an era where gerrymandering to protect incumbents is the norm, and campaign finance data from each state legislative race, mean that only a small share of these seats (probably less than ten or twelve a year) are seriously in play in any given year.

A couple of 300 likely voter surveys over the course of an election season in the ten or twelve most competitive state legislative seats would dramatically improve the accuracy of predictions of the post-election partisan divide of the state legislature, and this general approach could work in almost every state, as well as in elections for the U.S. House of Representatives.

Similar polling even a single time with similar sample sizes, even once over the course of an election year of local and regional ballot issues of consequence could likewise make a big difference to people interested in accurately predicting the probable nature of political landscape after an election.

There are less than 40 to 60 races for the U.S. House of Representatives that are remotely competitive in any given year, and given that these races are also conducted every two years, predominately involve incumbent candidates, can be analyzed with publicly available campaign finance data, and are held in highly gerrymandered seats, it isn't terribly difficult to quite accurately establish a list of viably contested seats. And, again, just two or three polls of a few hundred likely voters each conducted after primaries are complete in these races could dramatically increase the accuracy of predictions about the future partisan control of the U.S. House of Representatives, which is information of wide utility to a wide variety of people and organizations within and outside the United States. In principle, every single Congressional race has a potential significant impact on everyone in the world who is directly or indirectly impacted by federal government policy. The partisan control of Congress is almost as important, if not more so in many cases, on questions of national policy over an extremely broad range of issues, as who wins the Presidency.

In sum, conducting polling a couple of times in about 600 key candidate races each election cycle, and perhaps 300 polls an election cycle in notable regional and local ballot issue conducts, for a total investment of about 1500 polls with 300 likely voters each, involving less than 450,000 survey sessions of likely voters (less because some of the surveys could have partially overlapping samples, such as surveys of voters who live in both a surveyed state house and a surveyed state senate seat, which would be common since competitive seats are often geographically overlapping), plus additional calls to people who are ultimately determined to be outside the survey area, could dramatically improve the quality of the data available to political analysts. 

This might cost $1 million to $3 million per election cycle, which is a pittance next to the amount of money spent each election year in the United States on election campaigns, political analysis, and lobbying.  And, the highly geographically dispersed nature of this polling effort, as well as the fact that in our system of federalism no election crosses state lines (even Presidential elections are strictly speaking elections in each state for Presidential electors pledged to particular candidates), would facilitate cost sharing in exchange for co-branding the results when they are announced with the financial sponsors of these polls. Splitting the cost 50 ways into bundles for each state would bring the cost of this effort per state to something on the order of $20,000-$60,000 each per two year election cycle - something in the ballpark of what a single college or university in a state could afford (with some of funding with these kinds of sponsors coming "in kind" in higher educational institutions from social science student volunteers learning about research methods first hand, as several prominent colleges already do with great P.R. returns for the institutions involved).

This data would also make existing survey data during election years richer and more robust because the data when integrated with existing political polling, would tell us more about the political environment in which top of the ticket races are being conducts and would make it easier to identify outlier polling results in top of the ticket races.

In my view, this would be money well spent. Better polling of these races would also call more attention to them in a manner that would generally benefit the health of our democracy, because hose race coverage provides a natural starting point for media coverage of these political races and media coverage of political races increases public awareness of these races and improves the quality of the decisions made by the public in these races when the time comes for otherwise undecided voters in these races to vote.