On Colorado’s rigged debates system

Kyle Clark and 9News Denver NBC said that the reason the Libertarian candidate Lily Tang Williams and I weren’t in the U.S. Senate Debates were because we did not make the required 10%. Kyle said that we were asked to provide information if we had any polls proving otherwise. I asked what polls they were using and only received one. I sent this to Brandon Rittiman this information who was a moderator.

We weren’t asked to provide information. Polling does not ask about third party candidates. I was called and many others and we weren’t ask about Lily and I.

People under 34 aren’t polled. People without hardline aren’t called.

Third Party Candidates aren’t given equal air time on news so we aren’t known. So we have to go to extreme measures.
The debate last night did not use the words Global Warming. Both candidates were asked about fracking in a yes or no question and both said they were against local communities from banning fracking!

This debate nor neither candidate talked about endless wars, global warming, breaking up the big banks, auditing the defense budget, free college, not taking special interest money, ending the two party system run by the top 1%, arresting war criminals from our government, releasing whistleblowers, labeling our foods, getting universal health care, free early childcare.
It wasn’t a debate it was a commercial for the oligarchy to think we have choices that will represent the 99%. Here was the glitz and glamor between Bennet & Glenn by two repeaters not reporters.

Polling Discrepancies and Potential Errors

Here’s the research on polling:

There have been a total of 8 polls released on the U.S. Senate Race in Colorado. These polls are in ordered based on the most recent:

  • Emerson (9/9-9/13)
  • Magellan (8/29-8/31)
  • Quinnipiac (8/9-8/16)
  • NBC/Wall Street Journal/Marist College (8/4-8/10)
  • NBC/Wall Street Journal/Marist College (7/5-7/11)
  • Fox News (7/9-7/12)
  • Monmouth University (7/9-7/12)
  • Harper (7/7-7/9)

All of the above polls have raised questions about their methods and their reliability after review.

  • Emerson Poll (9/9-9/13) – The Emerson Poll is the most recent poll available. They polled 600 “Very Likely Voters” for the 2016 General Election. The statistics that stand out the most in this poll is that 593 respondents supposedly voted in the 2016 primaries (With 77 voting “Other” and 7 as “No Vote”). Based on the caucus and primary turnouts, these numbers would be hard to achieve unless the pollsters specifically screened respondents who only voted in the caucuses/primaries. Senator Sanders won the state by double-digits. The number polled for him 124 (20.7%) is lower than those who voted for Secretary Clinton 136 (22.7%). Since Republicans did not have a caucus or primary this year in Colorado, it is not known who would have won. Regardless, the results of this poll are invalid as none of the Republicans had a chance to caucus or primary.
  • The Emerson Poll shows Michael Bennet with a new low of 42.1% favorability rating. There also seems to be around 14.8% of the overall electorate that “Will Listen” to the other candidates for Senate. This poll also is the first since the beginning of July that has Bennet below 50% (Bennet 46.1%, Glenn 39%, Someone Else 6.9%, Undecided 8%). This poll did not include 3rd party candidates for the Senate race in Colorado.
  • Magellan (8/29-8/31) – The Magellan poll is one of the more detailed and reliable polls available as it has a complete demographic breakdown which others lack, and also include 3rd party options. This poll shows incumbent Senator Michael Bennet up 10% (Bennet 48%, Glenn 38%, Williams 4%, Menconi 3%, Undecided 7%). This poll leans Republican (which is possibly which it is not included in Real Clear Politics’ average). The respondents were 35% Republican, 34% Democrat, 29.5% Unaffiliated, 1% Libertarian, 0.5% Green). Arn is polling the best with single men (8.1%) and worst with married women (0.8%). Arn also polls the best with those 18-34 (7.2%), and support drops with age. CD1 has Arn’s strongest level of support at (8.1%). While questions can be raised about how conservatively skewed this poll may be, this poll may have the assumptions correct assuming more people turn out for Donald Trump over Hillary Clinton with the differences in enthusiasm in each campaign. Therefore, the Magellan poll may be one of the most accurate available.
  • Quinnipiac (8/9-8/16) – This poll demonstrates a lack of information regarding the demographics polled and does not publicly disclose their methods. This is a combination poll that also includes the Iowa Senate race. There is no full file available that shows anything about the type of people they polled, which makes it impossible to replicate. There are a total of 3 short paragraphs describing the Colorado Senate Race. This poll shows a 16-point lead for Michael Bennet, but fails to describe who is exactly voting for him. He is leading 54-38. Yet 8% is unaccounted for and no 3rd party options were described. This result is the same as the two NBC/WSJ/Marist polls conducted in August and July.
  • NBC/WSJ/Marist (7/5-7/11) (8/4-8/10) – These polls are among the strangest available and present numerous oddities and inconsistencies. There were no 3rd party options available in the CO Senate race, yet it does include 3rd parties in the presidential race. However, respondents were only given the choice of the 3rd party in the presidential race after being primed responding to the binary Clinton v. Trump question. No 3rd parties were asked about in the Senate race in Colorado. The only demographics described in both versions of the poll were male or female. An unusual aspect of the NBC/WSJ/Marist polls is that the pollsters were to guess the gender of the person they were speaking with. What jumps out as being very off with these polls is that every poll number for each question was different from July to August, except in the Colorado Senate race where not one number budged. The margin of error in these polls is around 3-3.5%. Therefore, receiving the exact same numbers (Bennet 53%, Glenn 38%, Other 2%, Undecided 7%) is quite statistically significant, and very improbable. No breakdown on demographic voting patterns was provided, even a gender analysis.
  • Fox News (7/9-7/12) – The Fox News poll skews towards the democrats and under samples independents (Democrats 38%, Republican 33%, Independent/Other 25%, Refused 4%). These numbers do not represent the electorate of Colorado. This poll did not include 3rd party candidates for the CO Senate race, but did include them in the Presidential race. Interestingly, Dr. Jill Stein is polling at 10% with Independents, the group most underrepresented in the sample. Additionally, this poll demonstrates a total of 17% either wouldn’t vote, do not know, or would vote for another candidate not listed of the major four. In the CO Senate race, Bennet is ahead by 15% (Bennet 51%, Glenn 36%, Other 1%, Wouldn’t Vote 3%, Don’t know 9%). Hispanics make up the largest portions of the “Don’t know” responses at 17%, followed by non-whites at 15%, and Independents at 13%. Collectively, Independents that “Don’t know” or “Wouldn’t vote” total 18%.
  • Monmouth University (7/9-7/12) – Monmouth is the first poll to ask about 3rd party candidates for U.S. Senate in Colorado. This poll shows Bennet with a 13-point lead (Bennet 48%, Glenn 35%, Williams 3%, Menconi 2%, Undecided 12%). Demographics were given, however they were not tabulated to include the portions given to each candidate. Independents were slightly under sampled (Republican 34%, Democrat 34%, Unaffiliated/Other 32%), and Hispanics were significantly under sampled (83% White, 3% Black, 11% Hispanic, 2% Other). More respondents said they would vote for Bennet (48%) and approve of him (51%) than found him favorable (43%). While it is expected to have some deviation in responses, the questions are correlated and the numbers would typically be closer together. This could demonstrate the straight-ticket voting mentality among voters and choosing the candidate who seems like the least terrible option.
  • Harper Polling (7/7-7/9) – The Harper poll is the first public poll available for the U.S. Senate race in Colorado. Democrats were slightly oversampled. This poll excluded 3rd party candidates and found Bennet with a 6-point lead over Glenn (Bennet 46%, Glenn 40%, Undecided 14%). Interestingly, the question was posed “Do you think Bennet has performed his job as U.S. Senator well enough to deserve re-election, or do you think it is time to give a new person a chance?” The responses were: Deserves re-election 38%, Give a new person a chance 38%, Not Sure, Depends on Opponent 24%. This piece of information shows a majority would be willing to vote in a new candidate depending on who it is (61%). Therefore, there could be at minimum 8% of Bennet’s 46% to defect to a better candidate. Only 20% of respondents viewed Bennet as “Very favorable” (Overall favorability 40%). This presents the possibility of up to 26% of Bennet’s supporters potentially being swayed even if they believe he deserves re-election. Oddly, somehow this survey includes 2% of respondents in the Albuquerque-Santa Fe, NM region. There is no data or summary showing they were excluded.

Poll Analysis and Summary

Multiple polls used a survey method called “Interactive Voice Response.” This is not person-to-person polling, but a form of automated polling which uses either voice recognition or touch-dial to select choices. Many callers may not be receptive to this method of polling and is often criticized from the poor design and response to these systems. Traditionally, polling has been fairly accurate to discern outcomes. However, polling certain populations remains a challenge and many polling outlets will weight responses they do not get enough of more heavily (i.e. making a 18-34 or Hispanic respondent’s answers count twice as much as more commonly polled individuals). Since all of these polls also ask about the presidential race, it is possible for respondents to be primed to answer a certain way once asked about the CO Senate race. It is expected to see polls with errors (as it is officially called margin-of-error) from the inability to sample a segment of the population. Yet it is extremely problematic that demographics and information tables were not provided in some of the polls that show the most potential for skewing and large errors in favor of a certain candidate. The NBC/WSJ/Marist polls have an anomaly of the same exact results that raises suspicions about the validity of their findings, and whether a second poll was even done in the month of August in the Colorado Senate race. The lack of demographic data on the NBC/WSJ/Marist makes it impossible to verify whether this is the case.

There are some concerns to be addressed that do appear to be real after an analysis of all these polls. It is highly unlikely, even with the problems within the polling, that Arn Menconi or Lily Tang Williams can reach a 10% threshold for the 9 News debate in the official “independent” polls. Highlighting this is the fact only 2 of the 8 polls (only one in the RCP Average) includes 3rd parties as an option, and both under polled segments of the population more likely to support a third party candidate. However, the real percentage points for both Arn and Lily are likely higher than what any of the two polls state, especially if considering a certain level of straight-ticket voting with the presidential election. Yet, part of the problem has to do with name recognition, which is correlated with Bennet’s high numbers. Additionally, Glenn’s consistently lower numbers are also a result of his lack of name recognition. A Senate race in CO has not been polled so heavily towards an incumbent Democrat in decades. Colorado has increasingly become a strongly progressive state with some political analysts and pundits removing it from the traditional “swing-state” category into the blue. Yet, the progressives in this state remain firmly divided into separate factions when it comes to the support of things like ColoradoCare.

There are signs to be optimistic that Arn can actually win this election, even when down in the single-digits at this point. There is no fear factor of voting the lesser of two-evils in the Senate of 100 and people should feel totally feel to vote for the candidate they think will best represent them. Colorado has shown that there are more true progressives here than moderate progressives that can potentially win the state for Arn. Yet, the Green Party must be careful about the strategy for winning Colorado as it includes also giving a boost to Lily-Tang Williams. There is promise that Lily can also generate a large following as the state has a lot of Libertarians that typically vote Republican. There is evidence that large factions of both Bennet and Glenn backers can be swayed to vote a different way. This can truly be a race where it ends something like 27% Menconi, 25% Bennet, 25% Glenn, 23% Williams. The recent primary with Bernie Sanders winning by almost 20% validates the fact that there are more strong progressives than moderate ones. Obama won the state with about 51.5% in 2012, with legalized cannabis beating him out by 55%. There is a definite possibly for Arn to win if the cards fall right.

Leave a Reply

Your email address will not be published. Required fields are marked *