Tag Archives: percents

The Election Pollsters Still Got It Right

election-forecastThere has been a great deal of Monday morning quarterbacking about how the 2016 Presidential election polls “got it all wrong.” Radio pundits like KFI’s John and Ken have been claiming that pollsters obviously don’t know what they are doing. There are three points to consider here.

1) Did the polls get it wrong?
2) Did the pollsters do something wrong?
3) What good math activity can we generate from all this fuss?

Here are some direct answers with, hopefully, simple, clarifying mathematical (not political) explanations.

The Polls Got It Right
The poll results were within the expected margin of error. In fact, four days before the election, Harry Enten of FiveThirtyEight wrote “Clinton’s lead is small enough that it wouldn’t take more than a normal amount of polling error to wipe the lead out and leave Trump the winner of the national popular vote.” In the end, Clinton still won the popular vote, by approximately 1.5% compared to the 3.3% predicated the day before the election, well within the normal margin of error. Gallup shows that, historically, the polls have been within 2%, on average, of the actual results, and within 1% half of the time, with the victories of Reagan in 1980 and Truman in 1948 being the most notable anomalies.

In fact, Nate Silver of FiveThirtyEight noted the day after the election that a 1% swing in Clinton’s favor across all states would have flipped the Electoral College tally.

Further support that the polls got it right comes from the understanding of probability. Clinton was given a 71% chance of winning on the eve of the election. That means that Trump had a slighter better chance of winning the election than he had of flipping heads on two consecutive tosses of a coin. When heads occurs twice when tossing a coin, should we all protest that statistics and polling are unreliable? This is why Nate Silver claims that the polls missed, but he did not say that they failed.

The Pollsters Did It Right
People have been willing to give more grace to the mathematics than to the mathematicians. Pollsters (those creating the polls, not the folks on the phone) have taken a great deal of heat for poor sampling, but these pollsters have been vindicated voter turnout numbers, because the pollsters surveyed registered voters, not guaranteed voters.

PBS‘s Michael Reagan writes that the data on actual casted votes reveals that Clinton had 2 million fewer voters than Obama did in 2012, while Trump had a slight uptick over Mitt Romney. Had voter participation been similar to the 2012 election, America would have had a different 2016 result.

Liberal filmmaker Michael Moore was extremely concerned just before the election about the lack of enthusiasm for Clinton versus the overwhelming passionate support for Trump. His concern turned out to be warranted.

A Good Math Activity: Secretary Clinton Attempts A Field Goal Kick
Given the information below from FiveThirtyEight, at what distance (in yards) would a field goal kicker in 2014 have the same chance of success as Secretary Clinton in the election of 2016.

election-percentage

Election Kickers.png

Spoiler alert: Approximately 48 yards.

Fortunately, if an NFL kicker misses a field goal attempt from just inside the 50 yard line, I still have faith in statistics and statisticians… and America.

Millionaires and Their Cars?

Car FerrariMy teenage son is preoccupied with three things these days: water polo, his girlfriend and expensive cars. He has fantastic talent in water polo, and has a wonderful girlfriend. He does not have an expensive car.

Currently, he is saving for his first car and knows that he will have to start with a used, low-end model, but he dreams big. He is always talking about Ferrari’s and Rolls-Royces. We like to talk about them together and point them out on the road whenever we are driving. He is convinced that he will own one someday. When I respond to his talk of grandeur, I want to sound like the Encouraging Dad (“Terrific, what kind of successful job do see you see yourself having, so you can afford that kind of car?”), but I worry that I sound like the Practical Dad (“That’s nice, it might be more realistic to set your sights on a cheaper car.”). The truth is that my words usually come out somewhere in the middle, which led to our very interesting math conversation the other day.

On a long drive back from a water polo game, we were talking about reasonable incomes (Practical Dad ruling the moment). He is a Junior in high school, so his interest is peaking about how much money is to be made as an adult. The conversation went like this:

Me: Guess what the average annual income is in America.

Preston: I know, because we talked about this in History class. $36,000 a year, but if I make $100,000 a year, a can save half of that and buy a rich car in five years.

Me: Do you know what percentage of Americans make over $100,000 a year?

Preston: 25%?

Me: It is actually about 4%. I know several people who make that kind of money. None of them drive a Ferrari, so you are going to have to make more than that. (Encouraging Dad trying to break through.)

Preston: If I made a million dollars a year, I could buy it in one year and still have enough to live on.

Me: With enough left over to care of me and your Mom. That would be awesome, but you are going to have to do something special, because less than one-half of one percent of Americans make a million dollars a year.

Preston: It has to be more than that. Look at how many rappers there are making bank.

Me: And think about how many are making just a normal living or how many are standing on a street corner singing while they hold their hat out for tips. Very few earn “Checks that look like phone numbers.”

Preston: Look at how many millionaires we know.

Me: I would say less than 5, off the top of my head.

Preston: Yeah, see?

Car LamborghiniInspired Math Question #1: If you know 5 millionaires, what percentage is that of all the people you know?

Inspired Math Question #2: If one-half of one percent of the people you know are millionaires, how many people would that be?

Preston: I bet there are at least a million millionaires in the country.

Inspired Math Question #3: Given that there are 300 million people in the U.S., and that 75% are adults, would one-half of one percent of American adults be more than a million people? (to be estimated while driving without a calculator)

Me: I am guessing that we are both correct on this one.

Preston: I still say it has to be more than that then. (Whether I am encouraging or practical, I am still Dad, so he must win!) Look at how many expensive cars we saw just today. There was a Ferrari, a Lamborghini and a Bentley.

Me: Yes, and think of how many other cars we saw today.

BentleyInspired Math Question #4: Approximately how many total cars might you see driving on a freeway for an hour on a Sunday afternoon? (must explain your reasoning on this one)

Inspired Math Question #5: If you see three expensive sport cars on that same trip, what percentage of all the cars would that be?

As we arrived home, Preston was still seeking victory. He is very good with mental math, so he knew where I was going with all the number crunching. In order to get the upper hand, he needed to bring in an expert, and what better expert in the world of teenagerdom to call upon than the internet? He Googled on his smart phone, “How many millionaires are in America?” and got an answer of over 3,000,000. He loudly reveled in glory. I countered with the age-old math argument of the importance of definitions. In this case, there was a difference between annual income and net worth. He was having no part of it. He was to busy flexing and bragging to Mom about how he just “owned” Dad in a math debate.