Jump to content
ExPDXer

Are we more different than similar?

Recommended Posts

After listening to one (too many) of MSNBC Steve Kornacki's psuedo-statistical political analysis, I got to thinking about demographics, groups, and DNA.
You may have heard Mr Kornacki,(or others) using phrases like 'college educated white females', or 'likely rural, middle class, asian american voters, over the age of 60, who identify themselves as evangelical republicans, and voted for Clinton in 2016'


Okay, you may not have heard that last one, but I use it as an example of the futility, inherent prejudice, and division caused by placing everyone into ever more detailed demographic boxes, then generalizing about their behavior for political, or marketing purposes.

 

Is this demographic data accurate? Useful?, Devisive?
 
I take some pride in being a member of the very smallest demographic. There is only one ExPDXer, and no statistician can predict how I will vote, no matter how much data they have on my race, religion, income level, geographic location.

 

Somewhere out there, two identical twin brothers, raised in the same household, voted for different candidates in 2016.

I believe that everyone is different. Very different, and unique. The minute you start putting humans into groups, and generalizing about their behavior, the accuracy of those predictions take a nosedive. This becomes dangerous when 'predictions' turn into assumptions, and stereotypes.


So what does this have to do with DNA?

At least 2 of the major demographic variables are DNA based. Gender, & Race. (There may be some Cambridge Analytica type group out there claiming to predict how people will vote based on hair color, but I digress).

 

This got me thinking, how many different DNA sequences are possible, and why do humans insist on putting themselves (and others) into groups?

Consider this:


Each human carries about 13,500 variants, and that perhaps 300 of these affect gene function.
The number of possible combinations is 1 x 2 x 3 … x 300 = 3 x 10^614. A number so large it can't easily be comprehended.

(There are currently about 7 billion living humans. The number that has ever lived is 108 billion.)
IOW, the human race has only used up a tiny fraction of the available combinations, only occasionally producing a Hawkings, or Einstein.

So we are very unique at birth, unique not only among others born in the same location, and place, but unique from anyone that has ever lived.
After that, a lifetime of different experiences creates further uniqueness. It's amazing that we agree on anything.
This may explain my Thanksgiving day political arguments with my siblings, and why twins may vote differently.

 

Animals herd into groups for two reasons, protection, and to accomplish large tasks.
It's impossible to communicate about politics without discussing various categories, or groups of people.
But, I think this always involves inaccurate generalizations, and therefore useless assumptions, and predictions about behavior.

 

 

 

Share this post


Link to post
Share on other sites
8 hours ago, ExPDXer said:

Somewhere out there, two identical twin brothers, raised in the same household, voted for different candidates in 2016.

 

Back in Canada, in junior college, 1974, I met a flamboyant, actually incredibly flamboyant, gay guy who had an identical twin. His twin went to a different community college. I met the twin a few years later. He was a straight, buttoned-down conservative business student. The gay one had a drug problem. The straight one, I heard later, completed an MBA and then studied law. I have no idea where either ended up.

 

You could easily confuse them, if the gay guy wasn't wearing makeup.

 

But that supports your point.

Share this post


Link to post
Share on other sites
20 hours ago, ExPDXer said:

Is this demographic data accurate? Useful?, Devisive?

 

Yes, yes, and yes.

 

I also like to believe I'm unique and my decisions are unrelated to my demographics.  I get hostile when someone tries to categorize me like that.  I'd be willing to bet that the vast majority of Americans feel the same way.  Another thing that galls me is when these analysts say something like "the candidate needs to reach out to the XYZ community".  That typically means talking to the self-appointed 'leaders' of these 'communities'.  I often disagree with the political positions that the 'leaders' of my 'communities' put forward and I resent the assumption that they speak for me.  

 

This analysis is divisive because it leads to prejudice.   People have always judged us by the way we look, the place we live, the car we drive, the clothes we wear...before they even say "hello".   All this demographic statistical analysis just reinforces that tendency to stereotype.  When the black lesbian social worker from Boston and the rural white Christian policeman from southern Alabama assume that they have no possibility of agreement, they're likely to approach any interaction with suspicion (at best) or maybe even hostility.  Just as bad, either of these people may be treated as traitors to their 'own kind' if they don't conform to expectations.

 

As much as I hate being put in a demographic bin, statistics show that people with similar demographic characteristics tend to make similar decisions.  The science of statistics isn't such that it's proven wrong when twins vote against each other.  The fact that I defy my demographic destiny on some issues does nothing to undermine the science.  It's accurate with percentages and probabilities over a population when using adequate sample sizes.  Statistical analysis of data comes with built-in measures of its own accuracy.   People tend to misinterpret probabilities and tendencies as absolute predictions.  

 

Demographic data helps a candidate tailor their message to specific audiences.  On the good side, it helps the candidate talk about what's important to that particular person.  On the bad side, it lets the candidate tell different stories to different groups.  Of course, candidates have always done that.   Demographic analysis allows a data-driven campaign to efficiently target their campaigning, sending exactly the right message to exactly the right neighborhoods to tip the election. 

Share this post


Link to post
Share on other sites
2 hours ago, Renegade said:

This analysis is divisive because it leads to prejudice.   People have always judged us by the way we look, the place we live, the car we drive, the clothes we wear...before they even say "hello".   All this demographic statistical analysis just reinforces that tendency to stereotype.  When the black lesbian social worker from Boston and the rural white Christian policeman from southern Alabama assume that they have no possibility of agreement, they're likely to approach any interaction with suspicion (at best) or maybe even hostility.  Just as bad, either of these people may be treated as traitors to their 'own kind' if they don't conform to expectations.

Yes. That’s a very good description of why this statistical stereotyping is dangerous, and divisive. It also influences individuals within these groups in a ‘peer pressure’ sort of way. And self reinforcing.

 

There is a very thin line between:

 

  • Predicting how a certain demographic will vote,

and

  • Suggesting (almost subliminally) how a certain demographic should vote.
2 hours ago, Renegade said:

As much as I hate being put in a demographic bin, statistics show that people with similar demographic characteristics tend to make similar decisions.  The science of statistics isn't such that it's proven wrong when twins vote against each other.  The fact that I defy my demographic destiny on some issues does nothing to undermine the science.  It's accurate with percentages and probabilities over a population when using adequate sample sizes.  Statistical analysis of data comes with built-in measures of its own accuracy.   People tend to misinterpret probabilities and tendencies as absolute predictions.  

 

I used identical twins to question the why certain DNA related attributes (like Gender, and Race) are even analyzed in the first place.

Why choose these variables in particular?

Wouldn’t variables like height, hair color, eye color, weight be equally valid or (equally ridiculous)? 

I’m sure given the data, that some correlation could be drawn between eye color and candidate preference, and even measured with a small margin of (sampling) error. It just seems like an arbitrary choice that was probably based simply upon the availability, and accessibility of a particular type of information, rather than it's influence on the item to be measured.

 

Accuracy… 2016 polls across the board were not accurate, IMHO. It least the bottom line conclusions they drew, and proclaimed loudly.

 

These polls consistently projected Clinton as defeating Trump. Election forecasters put Clinton’s probability of winning anywhere from 75-99%, and pegged her as the heavy favorite to win a number of states such as Pennsylvania and Wisconsin.

 

This is from the CNN right before the election 10/23/2016 :

 

New poll shows Clinton over Trump by double-digits

(CNN)Hillary Clinton has a 12-point lead over Donald Trump and has reached 50% support nationally among likely voters, a new ABC News tracking poll shows.The poll (with a margin of error of 3.5%) showed Clinton with 50% support to Trump's 38%, with 5% backing Libertarian Gary Johnson and 2% supporting the Green Party's Jill Stein.

 

The results showing Clinton with a growing lead largely match CNN's Poll of Polls, which averages recent national surveys. That Poll of Polls shows Clinton ahead 48% to Trump's 39%.

 

 

 

Okay, looks pretty good for Clinton

Actual result of the Popular Vote: Clinton won popular vote by just 2.1%.

So polls with a margin of error of 3.5% missed their prediction by 10%!!

This looks like an inaccurate prediction. Worse yet, it may have affected the election.

How many voters stayed home thinking that Clinton w/ a growing lead of 12 points, or 84% chance to win?

 

And from NYT:

poll1.jpg

This epic fail had polling companies like Pew scrambling for answers. They later produced a report that basically said the polls were not inaccurate, it was just a close election. A low MOE has a higher value to their clients, akin to gas mileage claims by automakers, your mileage will vary.

That just puts into question usefulness of a using polls with an average  +/- MOE  of 3%, to predict extremely modern election results with 1% margins. Results of 49.5% to 50.5% are commonplace recently.

 

But there are other problems…

  • Nonresponse bias- This occurs when certain kinds of people systematically do not respond to surveys.
  • Deception - Many of those who were polled simply were not honest about whom they intended to vote for. The idea of so-called “shy Trumpers” suggests that support for Trump was socially undesirable, and that his supporters were unwilling to admit their support to pollsters.
  • Identifying likely voters- Because we can’t know in advance who is actually going to vote, pollsters develop models predicting who is going to vote and what the electorate will look like on Election Day. This is a notoriously difficult task, and small differences in assumptions can produce sizable differences in election predictions. The Voters pollsters were expecting, particularly in the Midwestern and Rust Belt states were not the ones that showed up.

 

Understanding the MOE:

The margin of error is not the probability of the conclusion is correct. It used to be called margin of sampling error, and assumes that their method of conducting the poll is flawless.

In other words they analyzed the available data accurately, but it may the wrong data to analyze.

 

Over the years, MOE has typically been provided to give a sense of a poll’s overall accuracy.

That simple idea requires some critical assumptions, however:

  • It presumes that the sample was chosen completely at random,
  • That the entire population was available for sampling and that everyone sampled chose to participate in the survey.
  • It assumes that respondents understood the questions and that they answered in the desired way.
  • For pre-election surveys, it assumes that pollsters have accurately defined and selected the population of likely voters.

 

3 hours ago, Renegade said:

Demographic data helps a candidate tailor their message to specific audiences.  On the good side, it helps the candidate talk about what's important to that particular person.  On the bad side, it lets the candidate tell different stories to different groups.  Of course, candidates have always done that.   Demographic analysis allows a data-driven campaign to efficiently target their campaigning, sending exactly the right message to exactly the right neighborhoods to tip the election.  

 

  • It also helps candidates ignore statistically insignificant groups (minorities).
  • It also helps Russians, and others to micro target key demographics for disruptive interference.

I do not know whether demographic data indicated to HRC that should have spend more time in the rust belt neighborhood, or not.

The idea of a Presidential candidate targeting different messaging to different groups is repulsive, and disqualifying  to me.

I have a concern about the security of this data. It seems more likely that an adversary like Russia could hack, and possibly manipulate polling data to influence pubic opinion.

 

Then there's this:

 

At the direction of Trump, Michael Cohen paid a technology company to rig Trump’s standing in two online polls before the presidential campaign.

Cohen stiffed the owner of the technology company out of tens of thousands of dollars he promised for work that included using a computer script to enter fake votes for Trump in a 2015 poll of potential presidential candidates.

The company owner, John Gauger, told the newspaper that Cohen promised him $50,000 for the work but instead gave him a blue Walmart bag stuffed with between $12,000 and $13,000 in cash, plus a boxing glove Cohen claimed had been worn by a Brazilian mixed-martial arts fighter.

 

Share this post


Link to post
Share on other sites
38 minutes ago, ExPDXer said:

Then there's this:

 

At the direction of Trump, Michael Cohen paid a technology company to rig Trump’s standing in two online polls before the presidential campaign.

Cohen stiffed the owner of the technology company out of tens of thousands of dollars he promised for work that included using a computer script to enter fake votes for Trump in a 2015 poll of potential presidential candidates.

The company owner, John Gauger, told the newspaper that Cohen promised him $50,000 for the work but instead gave him a blue Walmart bag stuffed with between $12,000 and $13,000 in cash, plus a boxing glove Cohen claimed had been worn by a Brazilian mixed-martial arts fighter.

 

I don't know much about law, but isn't this a kind of fraud?

 

Wikipedia defines it this way:

"In law, fraud is intentional deception to secure unfair or unlawful gain, or to deprive a victim of a legal right."

 

Surely rigging a poll constitutes "intentional deception to secure unfair gain".

 

Does the fact that he paid with cash give Trump plausible deniability? As in "I had nothing to do with it, it was all Cohen, I knew nothing about it."

 

Since I know next to nothing about law, here's my worthless legal opinion: Trump ought to be impeached, convicted, tried in civil court, convicted again, then sentenced to years of hard time, with a possibility of parole after 85% of it. - And then be shanked 26 hours before he's eligible for parole. Call it "the McCabe conclusion".

 

Intentional cruelty must be met with cruel legal punishment.

Share this post


Link to post
Share on other sites

Since the outcomes of elections, in many cases, determine, which companies get more business and which get less;  And since polling results influence the outcomes of elections;  There is major incentive for big industries, like defense and oil, to impose their will on polling firms.

 

Most people are already aware, on some level that this is already happening, if not through direct bribes, then by the right or left leaning nature of the various polling firms.  Most Washington politicians, on either side of the aisle, take legalized bribes, which is reflected in the conflicting polling results by firms associated with one side or the other.

 

I would guess that the reason the difference is usually only a few points is because:  1 - Elections are usually very close and  2 - For purposes of credibility.

Share this post


Link to post
Share on other sites

 

2 hours ago, bludog said:

Since the outcomes of elections, in many cases, determine, which companies get more business and which get less;  And since polling results influence the outcomes of elections;  There is major incentive for big industries, like defense and oil, to impose their will on polling firms.

 

Most people are already aware, on some level that this is already happening, if not through direct bribes, then by the right or left leaning nature of the various polling firms. 

 

It’s not so much an issue of right/left leaning polling firms. Respectable firms sell data to both sides, and don’t particularly care about the content.

 

However, they are certainly an attractive target for corruption, and manipulation. Imagine if you could script the online polls to give your opponent a sizeable lead by placing fake votes. No candidate complains about being ahead, but it would depress turnout significantly. It’s easier to hack a central polling firm than a decentralized voting system, but almost as effective.

 

The basic problem with firms is that they are selling an inferior product of limited usefulness, and do not wish to admit it. They have a product that is ostensibly marketed to be accurate to +/-3%, but only have a total actual accuracy of +/- 9% after adding the errors of methodology.

 

This makes their product misleading, & only useful in elections with victory margins of greater than ~10%. They can correctly predict landslides. No one notices, or cares that their predictions were off significantly.

 

You would need an actual total accuracy of +/- 2%, be useful at all in predicting most close races. So none of these polls are even in the ballpark. But that will not stop anyone from consuming the stuff like junk food.

 

Statisticians working at these firms understand these errors, and usually provide very carefully worded statistical analysis or predictions, with abundant caveats, and disclaimers about what conclusions should be drawn.

 

Media outlets, and campaigns, and partisans take the carefully worded statistical summary (with disclaimers), and throw it in the trashcan, substituting their own partisan spin, and click bait headlines.

 

There are other bad actors, though. Cambridge Analytica openly claimed to be able to micro-target specific groups to change the outcome of elections by using data from Facebook, Experian, and other data consolidators.

 

They even have a name for election manipulation- "Persuasion DigitalMarketing", and news manipulation: "Persuasion Search Advertising"

Slides from Cambridge Analytica presentation:

1437.jpg?width=620&quality=85&auto=forma

 

Fake news

1437.jpg?width=620&quality=85&auto=forma

 

 

 

 

 

Share this post


Link to post
Share on other sites

Excellent topic and posts. Funny, today I received a letter that stated it was from the 15th congressional district in California where I live. The heading of the letter said it needed me to represent my district. I was curios so I read it. Inside it stated that they needed fair minded representatives to support Donald J. Trump.

My guess is that I received this letter because I am and have always been registered as an Independent. Seems like always, registered independent means you are right of center, or moderate. At least that's the way Independents are always characterized in polls and spoken about by the media. 

 

Actually, I've always been left of center. Much more so in the last couple of decades since it seems the only two parties worth speaking about have both moved towards the right. Today I guess they would characterize my views as leftist. Putting values aside, as it seems is what the two parties expect from voters these days I considered filling out the form and writing my own feelings regarding Donald Trump and the Republican party. But then I thought, that would be one huge waste of my  time. And it's no wonder people feel apathetic.

 

Peace!

Share this post


Link to post
Share on other sites

Putting what is in peoples DNA aside, people were a bit more aware in the old days. My own parents were depression era kids growing up. My dad was a registered Republican who actually stood in picket lines with unions that asked for better wages and safer working conditions. My dad was a conservative who was also not enamoured by JFK like my Irish Catholic Democratic mom was. I really cannot deduce anything DNA wise in that except to say political labels can and should be

taken with a grain of salt. Policy matters, it has to be clearly stated, along with the uphill battle that it might require. I say this, because I know that medicare for all is a long-term solution that makes sense for all. I've studied it and I am convinced that it can work. 

 

Policy matters and we need politicians who will speak the truth. GDP hit 3% in the first quarter and yes, oil prices are up. There are answers in the data and there are reasons why the government should break up Facebook. Yes, I am left of center, but still at the same time I have learned something from my old man who would be 94 today if he were still alive. 

 

Peace!

Share this post


Link to post
Share on other sites
On 4/27/2019 at 12:27 PM, ExPDXer said:

I used identical twins to question the why certain DNA related attributes (like Gender, and Race) are even analyzed in the first place.

Why choose these variables in particular?

Wouldn’t variables like height, hair color, eye color, weight be equally valid or (equally ridiculous)? 

I’m sure given the data, that some correlation could be drawn between eye color and candidate preference, and even measured with a small margin of (sampling) error. It just seems like an arbitrary choice that was probably based simply upon the availability, and accessibility of a particular type of information, rather than it's influence on the item to be measured.

 

Correlation isn't causation.  Even if eye color correlated to candidate preference, that wouldn't be enough.  There also needs to be some logical theory of how the two are related.

 

Yes, gender, race, and income are useful because that data (unlike eye color) is available by zip code.   But, it's the media that focus on gender, race, and income divisions.  They want something that looks like analysis, but is simple enough for the average channel surfer to understand.  Professional political analysts (e.g. Cambridge Analytica) are diving much deeper into the data.  The Wikipedia page gives a small taste of what that's like:

 

Quote

CA's data analysis methods were to a large degree based on the academic work of Michal Kosinski. In 2008, Kosinski had joined the Psychometrics Centre of Cambridge University where he then developed with his colleagues a profiling system using general online data, Facebook-likes, and smartphone data.[47][48] He showed that with a limited number of "likes", people can be analysed better than friends or relatives can do and that individual psychological targeting is a powerful tool to influence people.[47]

CA would collect data on voters using sources such as demographics, consumer behaviour, internet activity, and other public and private sources. According to The Guardian, CA used psychological data derived from millions of Facebook users, largely without users' permission or knowledge.[49] Another source of information was the "Cruz Crew" mobile app that tracked physical movements and contacts and according to the Associated Press, invaded personal data more than previous presidential campaign apps.[50]

Today in the United States we have somewhere close to four or five thousand data points on every individual ... So we model the personality of every adult across the United States, some 230 million people.

— Alexander Nix, chief executive of Cambridge Analytica, October 2016.[1]

 

OK, I just saw your later post about Cambridge Analytica.  Yes...that.

 

Have you seen those 'sponsored' quizzes at the bottom of web pages?  Yahoo is littered with them.  Take a quiz and give big brother a a few more data points to predict how you will respond to their advertising/influencing.

 

Let me argue the other side for a moment.  How is targeted advertising worse than general advertising?  It's all intended to influence voters.  Every campaign speech is an attempt to influence voters.   If I tell the AARP that I want to raise Social Security payouts and then I tell the Sierra Club that I want to reduce carbon emissions and then I tell my big donors that I want to eliminate red tape and simplify federal regulations (just making this up)...what wrong have I done?  If I do this in private meetings, it's politics as usual.  If I send computer-directed e-mails or if an algorithm makes certain ads pop up on specific computers...how is that worse?  It's not like I'm telling one group I'll repeal NAFTA and another group that I won't.  

 

On 4/27/2019 at 12:27 PM, ExPDXer said:

It also helps candidates ignore statistically insignificant groups (minorities).

 

What minority is statistically insignificant?  I'm not aware of any.  The whole point of the targeted advertising is that ever smaller groups can be made to feel that they're not ignored.  

 

On 4/27/2019 at 12:27 PM, ExPDXer said:

I do not know whether demographic data indicated to HRC that should have spend more time in the rust belt neighborhood, or not.

 

Either way, it doesn't reflect well on her.  Either she hired bad analysts, or she hired good analysts and didn't listen to them.

 

On 4/27/2019 at 12:27 PM, ExPDXer said:

The idea of a Presidential candidate targeting different messaging to different groups is repulsive, and disqualifying  to me.

 

Repulsive, yes.  But, I wouldn't say it's disqualifying.  They all do it.  I think it's just the nature of politics that, when speaking to a particular group, you emphasize your points of agreement.  They're just getting better at it now.  

 

On 4/27/2019 at 12:27 PM, ExPDXer said:

I have a concern about the security of this data. It seems more likely that an adversary like Russia could hack, and possibly manipulate polling data to influence pubic opinion.

 

My biggest concern isn't that polling data might be manipulated (I already assume that it is).  I still have hope that Americans don't vote a certain way based on a CNN poll.  Instead, I worry more about the online foreign agent trolls who try to stir up hatred and mistrust between different groups (Thousands attended protest organized by Russians on Facebook).   What happens when the Russians mobilize opposing groups to the same place at the same time?  Throw in a few covert agitators and things could get violent in a hurry.

Share this post


Link to post
Share on other sites
14 hours ago, Renegade said:

Correlation isn't causation.  Even if eye color correlated to candidate preference, that wouldn't be enough.  There also needs to be some logical theory of how the two are related.

Exactly. Causation, or some logical theory of what causes candidate, issue, or party preference.

 

I was searching for the logical theory that gender, or race causes them to vote a certain, aside from being told that they should fall in line their demographic stereotype. The actual results seem to harder to explain....

 

dem2.jpg

 

 

So I'm thinking, what is it that causes people to vote a particular way? Personally, I

Share this post


Link to post
Share on other sites

Continued...

dem2.jpg

 

dem1.jpg

Obviously there is correlation, but is this evidence of causation? What made non-college educated  white women vote for Trump 62% to 34% over a white women candidate?

Did  their gender cause them to vote for Trump? Was education the cause?

You might conclude that racial identity, is a stronger influence than gender, or education.

Maybe that's true. I hope not.

 

15 hours ago, Renegade said:

What minority is statistically insignificant?  I'm not aware of any.  The whole point of the targeted advertising is that ever smaller groups can be made to feel that they're not ignored.  

 

I see your point. This data is useful to campaigns for the reason you state. I was actually thinking about a blue neighborhoods in a ruby red zip codes, or a vice versa. I'm sure that battleground states will turn into battleground zip codes, then battleground neighborhoods, so maybe it's the opposite....talking to the trees instead of the forest. We need consistent macro, not micro-messaging that brings the country together, rather than divides along racial, or other lines. Also many are not represented in the polls as they do not have landline phones, or computers.

 

 

15 hours ago, Renegade said:
On 4/27/2019 at 1:27 PM, ExPDXer said:

The idea of a Presidential candidate targeting different messaging to different groups is repulsive, and disqualifying  to me.

 

Repulsive, yes.  But, I wouldn't say it's disqualifying.  They all do it.  I think it's just the nature of politics that, when speaking to a particular group, you emphasize your points of agreement.  They're just getting better at it now.   

 

If you are trying to deliver the same message to all voters, why would you need to segment your audience at all?

Emphasize, yes, but pandering, or adjusting your message to fit what pollsters think they want to hear is problematic.

 

 

Share this post


Link to post
Share on other sites
20 hours ago, ExPDXer said:

I was searching for the logical theory that gender, or race causes them to vote a certain, aside from being told that they should fall in line their demographic stereotype.

 

That's a good question.   If you ask Democrats (apparently that would include over 90% of black women) they'll tell you black women are rejecting the racist and misogynistic policies of the Republican party.  Of course, a Republican will tell you something else.  Although I couldn't tell you exactly what those racist and misogynistic policies are, until I have a better theory I'll tend to believe the people who are casting the votes when they tell me why they cast them.  

 

The social pressure you mention must also have an effect.  I wonder how many people say they voted a certain way in exit polls just because they don't want anyone to know they voted otherwise.  In some areas, you might not want anyone to know you voted against expectations.

 

18 hours ago, ExPDXer said:

I was actually thinking about a blue neighborhoods in a ruby red zip codes, or a vice versa. I'm sure that battleground states will turn into battleground zip codes, then battleground neighborhoods, so maybe it's the opposite....talking to the trees instead of the forest. We need consistent macro, not micro-messaging that brings the country together, rather than divides along racial, or other lines.

 

I agree.  But, I don't know how to make that happen.  Any ideas?  

 

This is a digression, but I think it's an example of the division you're talking about.  When the NFL players were kneeling during the national anthem, it made me angry that they were dividing America instead of uniting it.   The nation isn't perfect.  There's not a person alive who doesn't have some issue or grievance.  Should we all be making gestures during the anthem?  Global warming, deforestation, abortion, guns, free speech, taxes, inequality...the list is endless.  Instead of 50 different gestures for 50 different issues, why don't we all make a gesture together for making America better together?  Oh, yeah, we already had that:  Stand, take your hat off, and place your right hand over your heart.

 

The anthem should be a symbol of what we all aspire for our nation to be, not an opportunity to call attention to particular issues.  

 

19 hours ago, ExPDXer said:

If you are trying to deliver the same message to all voters, why would you need to segment your audience at all?

 

A politician's time with any particular voter is severely limited.  The voter will lose interest quickly, especially if the politician is talking about things that the voter doesn't care about, or worse, if it's something the voter disagrees with.  For the politician, it's only sensible to start with your best topic (for that particular voter).  If the voter's #1 issue is abortion and you just happen to have a position on abortion that the voter will like, then that's where you start.  If they only hear you say one thing, this is the one you want them to hear.  First impressions set the tone for everything that follows.  Then, the voter will be more likely to listen to, and respond favorably to, your second position.  As you go along, you build momentum.   Maybe you never get to the points the voter doesn't like.  But if you do, the voter is more likely to listen to your argument with an open mind because they know you have other points of agreement.

 

If you have to address a huge group of voters all at once, where do you start?  If you pick abortion, that's not what some want to hear.  Some won't agree with you on the economy.  Some won't agree with you on foreign policy.  No matter where you start, you're going to put someone off right from the beginning.  The beauty of targeted advertising is that it lets you put your best foot forward more often.

 

19 hours ago, ExPDXer said:

Emphasize, yes, but pandering, or adjusting your message to fit what pollsters think they want to hear is problematic.

 

I agree.  Politicians have been infamous for this since before computers even existed.  In some ways, targeted advertising enables this behavior.  But, there are risks.  If you send advertisements with opposing messages, it can hurt when you're found out.  Doesn't modern technology makes it more likely that pandering will be exposed?

Share this post


Link to post
Share on other sites
6 hours ago, Renegade said:

That's a good question.   If you ask Democrats (apparently that would include over 90% of black women) they'll tell you black women are rejecting the racist and misogynistic policies of the Republican party.  Of course, a Republican will tell you something else.  Although I couldn't tell you exactly what those racist and misogynistic policies are, until I have a better theory I'll tend to believe the people who are casting the votes when they tell me why they cast them.  

 

But they don’t tell you why. They are simply choosing a candidate from a list without explanation. Maybe there a poll out there asking that specific question, but it would have to be an open-ended question (not multiple choice), and therefore difficult to tabulate.  I suspect that the number of different reasons would approach the sample size of the survey, ranging from the logical (racism, economic policy, immigration policy, etc.) to the ridiculous (‘I don’t like his comb-over’, or ‘ I lost money in his casino’). I could not begin to write down all the possible reasons why anyone would choose candidate B over candidate A.

 

The most accurate polling data available is of course the election itself. With a sample size of 134 million, it only represents only 58% of the 250 million eligible voters, and even less (40%), of the total population of 327 million. Still, it is the benchmark for candidate preference surveys.

It is multiple choice, with no explanation given. In fact, no other data is available from this source except voting district, and party affiliation. They don’t disclose their education, income level, or anything else.

6 hours ago, Renegade said:

I agree.  But, I don't know how to make that happen.  Any ideas? 

 

I have no magic words. We can look back in history for some guidance. It seems like the only time the country is truly united in purpose is after national tragedies:

After the Great Depression, FDR made use of Radio technology to deliver his fireside chats. I believe this was useful in curtailing potential divisions within the country.

After Pearl Harbor, and during WWII, most of the economic, political, and racial divisions melted away, temporarily. Bringing the country together was imperative to winning the war, so the media focus, and political discourse was limited to war issues. It did not hurt that diverse populations from all demographics were brought together on the battlefield, and on the home front for common cause.

Of course there has to be a tragedy, and common cause. Vietnam did not bring us together, it drove us apart, because there was no unifying event, and no agreed upon common cause.

6 hours ago, Renegade said:

This is a digression, but I think it's an example of the division you're talking about.  When the NFL players were kneeling during the national anthem, it made me angry that they were dividing America instead of uniting it.   The nation isn't perfect.  There's not a person alive who doesn't have some issue or grievance.  Should we all be making gestures during the anthem?  Global warming, deforestation, abortion, guns, free speech, taxes, inequality...the list is endless.  Instead of 50 different gestures for 50 different issues, why don't we all make a gesture together for making America better together?  Oh, yeah, we already had that:  Stand, take your hat off, and place your right hand over your heart.

 

 Yes, there are as many grievances as there are people (even more, if you add my cantankerous uncle, who has an infinite number of them. Opinions flow out endlessly into the media, and the internet like a raging river. I do not expect agreement, I expect differences, it’s only logical. It’s harder to agree than it is to disagree, it takes work. Like I said, it’s amazing that we agree on anything.

Regarding the NFL, the anthem, and kneeling, I will defy my demographic by saying the political protest by players, coaches, or owners have no place during sporting events. Sports has a unifying purpose as a pastime, if politics stays away from it.

 However, it is also big business with employees. No other business that I am aware of starts their workday by playing the anthem, and requiring employees to stand as a condition of employment.

 

But perhaps I am unique in my views. I also think god has no place in the pledge of allegiance. I know others feel differently, but that's ok. Differences, and individuality should be expected. We don't need to fight over every issue.

6 hours ago, Renegade said:

The beauty of targeted advertising is that it lets you put your best foot forward more often.

Politicians have been infamous for this since before computers even existed.  In some ways, targeted advertising enables this behavior.  But, there are risks.  If you send advertisements with opposing messages, it can hurt when you're found out.  Doesn't modern technology makes it more likely that pandering will be exposed?

 

Demographic micro targeting, like any technology, has its bright, shiny side, and its dark hidden underbelly. I am not clever, or evil enough to tell you how this technology could be used for destructive, or corrupt purposes, but I'm sure someone is working on it.

 

You have made a good case for using this type of data in political campaigns, Renegade.

I accept this as current reality. I also suspect that campaigns want to understand the data fully, and use statisticians to gauge the accuracy, or lack thereof.

 

The media, on the other hand, does not want to understand the data. They want a headline. That usually involves reducing a complex data-set into simple graphics, and armchair statisticians opining, or coming to erroneous conclusions about what the data really means. I would be happy just to see error bars displayed on the graphs indicating the pollsters MoE, or even better, its real, absolute error as demonstrated by each polls historical performance.

 

Divisive? Yes. I think we agree. Not only promoting stereotyping, and prejudice, but also the rise of identity politics, where each demographic has a candidate designed specifically to meet their perceived preference.

Accurate? Not accurate enough to predict close races. Accurate enough to get a general ‘feel’ for where elections stand. No one should base their vote, or stay home because of this data. It’s like predicting the path of a hurricane. No matter how impressive the technology, it’s attempting to predict chaos, so there is a large cone of uncertainty. Feeling safe because you are outside of the thin line displayed on the TV screen is not wise.

Useful? Somewhat, accurate enough for basic campaign strategy. Useful as public information if it is presented properly.

 

It’s usefulness is diminished by it’s divisiveness, IMHO.

Share this post


Link to post
Share on other sites
On 4/30/2019 at 2:15 PM, ExPDXer said:

The media, on the other hand, does not want to understand the data. They want a headline. That usually involves reducing a complex data-set into simple graphics, and armchair statisticians opining, or coming to erroneous conclusions about what the data really means. I would be happy just to see error bars displayed on the graphs indicating the pollsters MoE, or even better, its real, absolute error as demonstrated by each polls historical performance.

 

Divisive? Yes. I think we agree. Not only promoting stereotyping, and prejudice, but also the rise of identity politics, where each demographic has a candidate designed specifically to meet their perceived preference.

Accurate? Not accurate enough to predict close races. Accurate enough to get a general ‘feel’ for where elections stand. No one should base their vote, or stay home because of this data. It’s like predicting the path of a hurricane. No matter how impressive the technology, it’s attempting to predict chaos, so there is a large cone of uncertainty. Feeling safe because you are outside of the thin line displayed on the TV screen is not wise.

Useful? Somewhat, accurate enough for basic campaign strategy. Useful as public information if it is presented properly.

 

It’s usefulness is diminished by it’s divisiveness, IMHO.

 

Big data is interesting in that you can use it to focus on facts in certain areas or demographics. Then, comes the why. I think that's where you kind of have to put on your hard hats, because the why is usually not as straight-forward as anyone would like it to be. Human beings are incredibly smart, they come from families and from a whole host of social affairs that can diverge in almost every individual family. Why was the middle kid the one who excelled, or the one who became the black sheep of the family? The data points to statistics but does not necessarily answer the question, nor does it indicate the odd stat that does not fit the mold of the big data. 

 

Yes, we generalize. Everybody does in some way generalize. I think that is part of human nature. So you have to lay out specific qualities that are not in themselves political in nature. 

Some uncle wears a red MAGA hat and complains about the fake news, and yet still he is bothered about how inequality has affected his own life. Possibly he thinks it is because there is too much government regulation, and that people on the left all got together to pull a hoax on the populace regarding climate change. He reads into all the news that somehow there is this giant conspiracy that has been brought about by a elite and powerful fake science left. But he's not alone, there are others who feel the same. 

 

Is it in his DNA, was he born to be like he is? That's a million dollar question. Actually, I would say that's the trillion dollar question these days. 

 

I'm not a social scientist, nor anybody who could sort through the big data and from it deduce important conclusions. I not even a smart guy. However, the old adage of a tangled web that has been sewn, to me does indicate a big part of the problem. Take it from an idiot. Sometimes being objective, is not always easy. 

 

Peace!

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...