NFL Players Taking a Knee is More Complex and Polarizing than We Think

Tom H. C. Anderson
October 1st, 2017

If a topic is worth quantifying – it’s also worth understanding the ‘Why’s’ behind it

I think I’ve seen at least 10 polls on the NFL Protest issue in the last week. As usual, these polls are simple structured questions, sadly usually not much more than a simple agree or disagree option. For example, the CNN poll released this weekend showing 49% Against 43% Supporting, (I’m not quite sure what happened to the remaining 8%). A PBS/Marist Poll had 48% of Americans saying protests were respectful and 46% disrespectful. Yet another, Seaton Hall Poll, had 84% supporting the players right to protest!

Which of these is correct??? It seems with structured questions, you really do Receive what you Ask for

It’s different when you allow people to answer whatever they want. As we have done here on the blog several times before, in order to challenge structured/forced-answer polling and explore the results we get when people are allowed to say anything they want about a topic. This weekend we asked 1,500 Americans 1 single question:

“Q. What are your thoughts about NFL Players ‘Taking a Knee’ during the national anthem? [Please elaborate why you are for or against what these players are doing]”

We utilized our text analytics software platform OdinText to classify responses into 3 groups. Those clearly against the players protesting (46%), those supporting the players taking a knee (33%), and those who did not take a clear position either way, i.e. they either did not care, understand, or had more mixed emotions on the topic (21%). We then weighted these groups based on location, age, and gender to as accurately as possible project onto the general US population. Post weighting comments fell out slightly more even, though still with a majority Against the protests (43%), over a third Supporting player protests (37%) and one in five not taking a side explicitly (21%).

NFL OdinText Chart 1

In past text analytics polls, we have seen differences from other mainstream structured polls which seem to indicate text polls can provide slightly different, more accurate proportions as they relate to actual behavior (our pre-election TextPoll indicated the Trump-Clinton upset ahead of time).

We believe this may partly be due to the fact that when you ask someone why they feel a certain way, there is additional cognition when answering. Of course the true beauty of text analytics lies in better understanding and quantifying the relative importance of the various Why’s.

The WHY’s Behind The Opinions

Visualizing the comment data about WHY Americans actually feel the way they do about the protests quickly paints a picture of the sentiment and the reasons behind them.

OdinText nfl vizualization 2

Along the X axis from left (those against the protests in red) to right (those supporting the protests in green), with Y axis representing frequency of mentions, we see that “disrespectfulness” is the most frequently occurring theme, and it is indeed more frequently mentioned than the primary reason for supporting the protests on the right in green “freedom of speech”.

In the chart below we can see the most important reasons WHY fans agree or disagree with player protests even more clearly (sorted by difference gap).

NFL OdinText chart 3

Beyond looking at sentiment and what was answered, with text analytics we may even consider the mechanics of how questions were answered. Those who were not clearly for one side or the other, while answering fastest (49 seconds on average) and having shorter answers on average tended to use slightly longer/complex words. Those Against, while taking longest to answer (1.5 min on average) tended to have responses of more similar sentence and word length as those Supporting the player protests (though the latter group answered somewhat faster, in about 1 minute). This longer response time might give some indication on the relative importance of the issue to the group, i.e. those Against the protesters may take the issue more seriously/personally.

The most interesting thing here, though of course, are the specific topics mentioned by the two groups. Among those Against the players taking a knee, Disrespect for flag, Country and Armed service members was the most frequent issue mentioned. Several mentioned that they would be choosing not to watch Football/NFL. Among respondents who self identified as Veterans in their response, somewhat more were in the Against camp.

Freedom of Speech, and the Right to Peaceful Protest were the two most mentioned topics related to Support for player protests.

The two sides seem to differ most on the appropriateness of venue, with many of those disagreeing with the players, pointing out that they are well paid (over paid in fact), and that the work place/NFL/Entertainment is not the place for politics. The idea of appropriateness not just of venue, but respect for flag and country was also brought up frequently by those with more Mixed opinions.

Don’t Forget to Drill Down Some

I would be remiss if I told you three simple charts were sufficient in order to understand the American Psyche on an emotionally charged issue such as this one. One of our users pointed this out quite elegantly in a guest post recently here.

While many times a good visualization or two are enough to communicate everything you need to know in a text analysis, other times, we do need to remind ourselves to give a few actual example quotes.

While obviously not in our data for this study, I believe Obama may have said things well on the NFL Protests before leaving office “I believe that us honoring our flag and our anthem is part of what binds us together as a nation. But I also always try to remind folks that part of what makes this country special is that we respect people’s rights to have a different opinion…it is important for everybody to listen to each other. I want (the protesters) to listen to the pain that that may cause somebody who, for example, had a spouse or a child who was killed in combat and why it hurts them to see somebody not standing… But I also want people to think about the pain he may be expressing about somebody who’s lost a loved one that they think was unfairly shot.”

I’m leaving you below with about a dozen sample quotes (more than I usually would recommend) for each of the three groups we identified above [sic]. These days, perhaps more so than ever before, I think we really do need to take the time to listen to each other. I welcome your thoughts to this post below, and am curious if any of the specific example quotes below resonate with you and if so which one?


“stupid disrespectful to our soldiers these are paid men just do your job what if doctors and nurses stated their political views when you where going into surgery a person wouldn’t care for that either”

“I’m completely against it as an immigrant i have always been in awe of america’s patriotism in this political climate we need to stand strong together and love our country otherwise this country will not last”

“i think it disrespects the u.s the national anthem an the american flag and is an inappropriate expression of private political views in the workplace during the player’s performance of his duties and disrespects the fact that the fans in the stands have paid their hard earned money to watch a football game not to be subjected to the political views of the players the fans aren’t paying for their tickets so that players can express themselves at work”

“im all for free speech but not during our national anthem i find it completely disrespectful for the true heroes that chose not to sit down and fought for our country there is no debate it’s classless and demeaning and a distraction to sports”

“I am not allowed to make political statements at my job they shouldn’t at theirs either the NFL wouldn’t let the cowboys wear a patch in support of police yet it will allow this madness I am done with the NFL for life”

“Just play football I don’t tune in to have politics pushed in my face if i wanted to hear others political views I’d watch cnn not the NFL”

“wWhy not protest any cause that anyone thinks of, valid or not, have no order at public events everything wrong in the world should be protested not just selected causes not fair to just limit to a few causes”

“As a veteran I am completely against anyone who takes a knee or protests during our national anthem I understand people feeling a need to peacefully protest but during our national anthem is not the time”

“Counterproductive and self indulgent inevitably perceived as a poke in the eye to the USA protest doesn’t illuminate the cause hence a failure”

“They have a right to protest I have a right to not watch and cancel my NFL package which I have done”

“iIm against it they need to focus on football and not do something that further puts our president in the spotlight they’re only rewarding the very person they’re trying to speak out against”


“Stupid controversy desperate president grasping for distractions”

“I don’t believe I could care any less about NFL nor the national anthem and how to listen to it…”

“I am very mixed on this. Part of me thinks it is wrong and getting out of hand that our flag and national anthem represent the freedom we have from sacrifices of those who have fought and died to protect us and that kneeling represents lack of respect, but like it or not the freedom allowed to live in our democracy gives us the right to do things as long as people are not physically hurt or mentally crushed freedom to protest i cannot give better answer”

“There is a political process…this…and Mr President please stop using twitter and calling names that is not appropriate either it all causes strife… players want to make a statement then use the process”

“I am for their freedom of protest I am against using the NFL as a venue of protest leave politics out of the NFL”

“People have the right to express themselves people fought for the right to peacefully protest no one is against the us army or the flag sports and politics shouldn’t mix either have players stop coming out of the locker room for the anthem or stop doing it all together at games”

“They certainly have the right to protest however sports does not need to be mixed with politics sports is an escape from that”

“I’m neither for or against the players what i am against is the president and his endless cruel remarks on any subject”

“As long as they decide what to do as a team who cares. Before the year 2000 players were never on the field for it anyway”

“I am a veteran part of me accepts that one of the things I left my family and home to do was to protect others freedom of speech on the other hand I think there are ways of protesting that don’t show disrespect to the nation via its symbol the flag that gives you those rights and doesn’t serve as a slap in the face to everyone who has saluted it fold it to hand to a devastated family member or had one of those folded flags handed to them”

“I fully support their 1st amendment right to do so and think it is absolutely ridiculous that they are expressing their outrage against a country whose unparalleled and unprecedented levels of personal freedom and protection have made it possible for them to become multi millionaires among 1 of the wealthiest people on earth in their 20s for playing a game the right to free speech means the right to say and support ridiculous things including ridiculously myopic grievances they should try living one month in the life of any ordinary citizen from basically anywhere in the southern hemisphere and see if they ever complain again about life in the united states but go ahead kneel sit lay down stand on your head whatever makes you feel”


“They began protesting the plight in america of African Americans now the president has forced them to protest him it’s not illegal to not stand hand on heart for the american anthem and it is their constitutional right to protest ills in our country any way they choose no one mentions but part of every oath for public office including president say you will defend the constitution the president needs to read it”

“We have freedom of speech period not being able to exercise that makes us no different than Russia or North Korea or others exercising our freedoms should be the highest honor to the military why else have they fought and died plus the flag represents America not the military not veterans the national anthem represents America not just the military and veterans the flag flies over government buildings it’s carried by olympic teams it’s hung on lawns it’s a symbol of America and does not solely belong to patriotism or the military or veterans by requiring them to stand would be forced nationalism and the first step to losing our freedoms that veterans have fought for to require them to stand would be spitting in the face of those who died so that we can be truly free”

“I think they are courageous to use their position to express their concern about inequality in America as both the athletes and many veterans have expressed the kneeling is not about disrespect to the flag or the anthem sad that we must disagree about everything these days it seems that people would rather believe it is an act of disrespect than even try to understand the issue”

“I applaud it you don’t have to agree with their stance to be glad to see people exercising their rights to free speech”

“It’s an effective public protest against a president that solely focuses on public image at the cost of any substantial or meaningful actions”

“I’m for it but it should be for police brutality against all Americans it is not just blacks that are being harmed”

“It is their freedom of speech if we force people to stand we are no better than North Korea, China or Syria all places that force standing to their pledge”

“I think their stance is admirable and courageous taking a stand for those who have no voice”

“Ok they are like it or not role models and seem to be on justice’s side and if the pres would shut up more we would all be better off”

“I am for the nflNFL players taking a knee our flag means nothing if your brother is not included or does not feel included is mistreated or feels mistreated we should address the needs of all citizens and not just a few”

“Only black men with a consistent platform to make a statement regarding racial inequality”

“It’s their right this country takes everything too seriously the citizens of this country need to relax”

“Taking a knee is what sports players do to show respect when a fellow player is injured it is another way to show respect and courage”

“They are right to use their position in the spotlight to draw attention to social issues within our society the issues at hand are not about respecting our veterans or the flag itself and we should not be distracted by these arguments”

“You can’t demand respect”

“It’s their right why are public displays of patriotism necessary in the first place”


NOTE: This TextPollTM is a quick non client project intended to demonstrate unstructured data analysis using the OdinText analytics software. That said, results have a confidence interval of +/- 2.5% at the 95% confidence level (greater than the 3 outside polls mentioned at beginning of post). If this had been an actual study we would have recommended looking at data by political affiliation as well as NFL viewing.

About Tom H. C. Anderson

Tom H. C. Anderson is the founder and managing partner of OdinText, a venture-backed firm based in Stamford, CT whose patented SaaS analytics platform is used by companies like Disney and Coca-Cola to mine insights from complex, unstructured and mixed data. A recognized authority and pioneer in the field of text analytics with more than two decades of experience in market research, Anderson is the recipient of numerous awards for innovation from industry associations such as CASRO, ESOMAR and the ARF. He was named one of the “Four under 40” market research leaders by the American Marketing Association in 2010. He tweets under the handle @tomhcanderson. Click here to request a demo or additional info.








16 thoughts on “NFL Players Taking a Knee is More Complex and Polarizing than We Think”

  1. Hi Tom! Permit me to say that this is a very good example of qualitative methodology, with a sense of proportion included. It demonstrates to those who would like to know more about the thinking and especially the emotions behind consumer behaviors how they might acquire that knowledge.

    In the setup for your review of your findings, you ask: “Which of these is correct??? It seems with structured questions, you really do Receive what you Ask for…” You then go on to report your findings on…

    ““Q. What are your thoughts about NFL Players ‘Taking a Knee’ during the national anthem? [Please elaborate why you are for or against what these players are doing]”” The pie chart suggests that your survey actually measures the For/Against proportions of the population you sampled, as would a typical quant study. That implication bothers me, as do almost every study whose findings imply representation of a populations.

    In today’s world, I suspect any news story or pundit that says ” Americans do/don’t (fill-in the rest).” Even when the report in the news/print offers a note on how the sample was defined, weighted, data-collection method(s), I am skeptical of the representativeness of the sample. It is not enough to claim that all studies are subject to the same limitations, or there is no known way to overcome non-response. When The News claims this or that about a population, the organization is influencing the public, just as a market researcher is influencing a business.

  2. Thanks Terry. The data may be unstructured “Qualitative” as you say, but text analysis most definitely is quantitative. The explanation of how the groupings were done is there along with example quotes from each group at the bottom. I think these 3 groups made most sense given the question and data, those clearly for, those clearly against, and a third group comprising of those who do not fit into the first two.
    I take offense with the ‘new’ thinking that research should somehow be politically correct and not report findings in fear of ‘influencing’ a discussion. Nonsense IMO. We are following scientific process and applying best statistical practices. While I’m typically not in favor of weighting, because this was done using Google Consumer Surveys and they suggest weighting schema to gen pop we applied it. Typically their weighting doesn’t change the results much at all, here it did change the numbers slightly, so I chose to report it.
    I agree with you in terms of quality of the typical sample firms and that’s one reason I think text analytics is great. You can tell by the comments that these are real Americans who are familiar with the issue, and not a ‘survey farmer’ from some other country gaming incentives. One of the reasons we have used Google so many times to do these example analysis is exactly because of the quality (they put themselves between real valuable content on the web from the WSJ to movies and music and have a lot of additional data on each respondent), not to mention they are also fast and inexpensive. But we have also partnered with some of the better panel companies and will continue to do so in areas that don’t make sense for simple Gen Pop sample.

  3. Very interesting. I find myself wondering about the implications here. Is it possible those who disagree with the protest are viewers or perhaps heavier viewers, and those who agree with the protests are not, or are lighter viewers? I kind of read that into the comments (those against protest mention purpose of game/viewing while those who agree seem to focus on rights), wondering if you asked?

  4. Absolutely Dale, Would recommend looking into the difference between NFL viewers here if you were NFL or an NFL partner or advertiser. However yes, I suspect you are right, and as you can see in the bar chart, second bar from top, those against the protests were far more likely to mention Football and the NFL. So likely to be a bigger problem for them.

  5. Tom, I appreciate your response , and accept your reasoning.

    I definitely believe that the Media outlets of all sizes and stripes use survey results from any source to “report” news that fits their own agenda, increasingly with a political bias. To aid their purposes there is no shortage of survey purveyors to feed them the results of studies that fill the bill.

    Ironic that you should mention Google/WSJ, as I am in pursuit of the details of an NBC/WSJsurvey reported on Meet the Press recently. The Data Download that is reported on Meet the Press is a Prime Example of my pet peeve!

  6. What’s Ironic these days is I see researchers being attacked for posting neutral completely clinical analysis, simply because readers don’t like the results and/oe they portray their position as less popular. It’s akin to getting angry at your doctor when she/he gives you a diagnosis. Crazy!

  7. For the polls you cite, hopefully, the sample selection, research methodology and weighting procedures are all comparable. In Canada for example, companies use different methodologies – IVR , online and sometimes telephone still.

    The question metrics you cite by the companies though are quite different (against/supporting, disrespectful/respectful, right …) and the question wording will likely have been different as well.

    It would seem that of the companies’ you list your company’s question metric is the same as CNN’s. It’s further interesting to note that the against less supporting percentages at six points are the same. The higher “not sure, both against and supporting” percentage based on the Odin poll may suggest that certain people are a bit more ambivalent when allowed to take the time to express an opinion open-ended than would be the case closed.

    Since most polls are closed-ended, the Odin “not sure …” percentage compared to CNN’s does beg the question as to how accurate closed-ended polls are. CNN’s might be an easier story to tell for media. Odin’s you could add the not sure percentage to either against or supporting or disregard them and recalculate.

    Personally, I prefer your story … and the diagnostics.

  8. Hi Tom. I have reviewed a number of your text analytic applications and as a researcher, I appreciate this rich form data gathering and quantification. Being able to read respondents’ typed verbatim comments is very helpful for context. It is unfortunate that our practice does get misused and abused for political and other gain.

    I expect you give appropriate consideration for sampling in your research. Unless an entire population is surveyed, there will always be some amount of potential bias due to sampling. That is inherent in research. As professional researchers, we balance minimizing that with cost and time. No further comments on your research technique.

    To offer my perspective on the topic, I don’t see why the National Anthem should be played at US sporting events. I’ve read the history but I do not see the value of playing the National Anthem at US sporting events (two US teams sports teams playing each other) – the place of work for the sports players. Yet employers and employees for all other types of professions aren’t required to listen to the National Anthem every day. I say discontinue the National Anthem at sporting events unless the events are US and international and other countries play their anthem.

    Standing or not standing for the National Anthem is really much ado about nothing (in my opinion). What is important, now that these events have occurred, is to understand WHY. This is the point that so many citizens are missing including our President (or do not want to acknowledge). If this country wants to disregard or change our constitution in terms of the freedom of speech, then let’s also change (do away with) the right to bare arms. No civilian should have the right to own the type of weapons used in the unthinkable act that just happened in Vegas. My thoughts and prayers are with those impacted.

    Tom, I enjoy your Odin Text topic reviews. I’d like to see more analysis on the increasing divisiveness being displayed in this country. It is very concerning.
    Keep ’em coming!

  9. Thanks Mike!
    Thanks Lesa. Anyone can use OdinText to do these types of analysis with any data of their own. I do like using real data to demonstrate with. We’ll definitely keep em coming!

  10. Tom, despite the irony, and frankly I’m not sure how I “feel” about the numbers in the pie chart, I prefer to believe as in a matter of faith, that surveys including/especially verbatim response driven surveys have an important and beneficial role to play in industries where products and services are consumer oriented, and in B2B.

    Arguments over whether or not voter/public issue polling is in fact a consumer based undertaking aside, I challenge the merits, and fear the consequences, of polling political issues.

    I applaud your tool, but I would only use it for topics as in the case in point for qualitative learning. For product and service applications I think you are ahead of the research industry and offer a unique and valuable quant research alternative. I attempted to accomplish a one-off trial of a verbatim only quant study in the early 1990’s. While I was far from satisfied with the outcome, I’ve never given up on the concept. Hence I would readily recommend that a client firm or researcher use Odin Text in a commercial context, i.e. not sociopolitical.

  11. Excellent work – as a professional market researcher for over 20 years at large companies your approach combines the best of open ended questions while obtaining closed ended like results – but sampling and projections are key – my issue is projecting to total populations wirhout factoring in “heaviness of use ” in this case I recommend reporting and possibly weighting not just by age, gender, region etc. But by:
    1. Political affiliation – republican, democrat, independent
    2. NFL fans vs. not –
    3. # of NFL games watched
    4. Whether attended a NFL game live or not
    Having worked on large categories like Soda – which conducted many forced choice preference studies between Pepsi and Coke – I found that age and demographic weirghting was insufficient leading to unstable results and not until prior brand preferences were incorporated the projections were inaccurate and unstable

  12. Thank you Arthur. Absolutely. Had this been a project undertaken for NFL I would have suggested that those two variables should be considered. Again, these were used because they are provided free of charge by google, we only asked the one single question, which is the cool thing about this approach.
    That said though, this issue is relevant at the national level and awareness is mainstream. Case in point, these days I consider myself left leaning independent, and never watch a football game willingly, but sadly, until today I couldn’t seem to get away from the issue. So I do think if we want to understand impact at the national level, the additional two variables you mentioned are less important. If we wanted to understand advertising impact, then you are absolutely correct.

  13. Tom, check me on this: What I get out of the verbatims and themes is that those against the kneeling protest are against for two dominant reasons:
    1) They are disturbed to have the issue brought up during their fun time, because the protest highlights their cognitive dissonance between “the U.S. is great” and “injustice is being done”. They didn’t tune in / pay their cable bill to be disturbed.
    2) They accept the narrative that the protest is somehow anti-military.
    This leads me to the further conclusions that the protest successfully delivered its message, but can’t deliver sympathy with that message; and that it would be interesting to examine the origin of that narrative (2).

  14. Ian, I think you could look at this a number of ways. At the very highest simplest level though, as you can see in the second chart/visualization I posted, the single most frequently mentioned theme was that this is “disrespectful” to the flag, country, soldiers whatever. This was mentioned even by some who might have also mentioned some sympathy with either their right/freedom to not have to stand and/or the cause itself.
    What we have here is a mixing of more than one issue. I don’t know that it was initially thought out super well for maximum positive impact for BLM. Most certainly in generating awareness it has been successful, but the awareness of not standing is far higher than mentions of the core issue intended to be protested.
    We didn’t ask any other questions other than that 1 OE, as the intention here is to demo software. However it seems to me that NFL fans are probably at least as patriotic or probably more patriotic than your average American. And this is a weakness of the protest and potential problem for the NFL, one that Trump may have seen and decided to exploit.

  15. Hi Tom,

    Hope all is well. Not sure if I will be making it to TMRE this year. If I do I hope that there is an opportunity for us to catch up.

    Just a few thoughts about your observations.

    At Gadd we do more qual projects but in the past year our revenue has been higher for quant. Whenever we can we like to include open-ended questions in our quant. Our qual though is much deeper understanding “why” (laddering to “how”) insight driven.

    For the polls you cite, hopefully, the sample selection, research methodology and weighting procedures are all comparable. In Canada for example, companies use different methodologies – IVR , online and sometimes telephone still.

    The question metrics you cite by the companies though are quite different (against/supporting, disrespectful/respectful, right …) and the question wording will likely have been different as well.

    It would seem that of the companies’ you list your company’s question metric is the same as CNN’s. It’s further interesting to note that the against less supporting percentages at six points are the same. The higher “not sure, both against and supporting” percentage based on the Odin poll may suggest that certain people are a bit more ambivalent when allowed to take the time to express an opinion open-ended than would be the case closed.

    Since most polls are closed-ended, the Odin “not sure …” percentage compared to CNN’s does beg the question as to how accurate closed-ended polls are. CNN’s might be an easier story to tell for media. Odin’s you could add the not sure percentage to either against or supporting or disregard them and recalculate.

    Personally, I prefer your story … and the diagnostics.

Comments are closed.