Data Management: From Politics to Poker

With elections in the recent past, and opinion polls cropping up at every turn telling us to vote for one candidate or another, it is important to remember one case that shows how these polls can sometimes be misleading.

The 1936 elections pitted Republican candidate Alfred Landon against the now-famous and then-president Franklin D. Roosevelt. A magazine called The Literary Digest put together a mailing list of a whopping 10 million subjects for their presidential poll, out of a number of readily available lists, including magazine subscribers, car owners, and for the first time ever, phone users. The magazine had, at the time, been famous for its accurate prediction of the past four presidential elections in a row, and so when the poll returned a prediction of Alfred Landon winning the election with 57% of the vote, readers of the magazine had no reason to doubt it.

When the elections came along, however, Roosevelt crushed Landon in a landslide, carrying 62% of the vote behind him. This 19% error is the largest ever recorded for a major public opinion poll, and a massive departure from the accuracy that the magazine had boasted in the past. So what caused this enormous change?

The answer lies in the switch to using telephone lists in order to get a larger sample size. Although it seemed that a larger sample size would make the survey more accurate, in reality it only skewed the survey towards favoring those who were more wealthy, since they were the people who would have had a phone in their home at the time. Although a tragedy for the publication, it is now one of the most famous examples of selection bias to have ever been recorded.

At WCI, there is a course that teaches about these kinds of biases, and how to avoid them. The course Data Management is taught by Ms. Funk who describes it as “a probability and statistics based course, using probability to predict future events or using statistics to describe things that have already happened”.

In her eyes, the course is useful for a variety of different fields. Students going into business, economics, politics, or any number of other paths that involve the use of statistics even in the smallest of ways will often have to take a statistics course in their first year of post-secondary education, and taking a similar course in high school would give those people a leg up in understanding the course material.

The course also has a number of practical uses in everyday life. We are all exposed to a variety of statistics and data in the media every day already, Funk suggested, and Data Management’s units on bias and sampling are there to help students identify and overcome biases, not only in their own work, but in the work of others as well. Lucas C., who is currently taking the course, agreed. Lucas intends to study political science, and he said that “even from a general standpoint, it’s important that people be better equipped to interpret and consider different options that are presented to them in their life.”

Michael F, who took the course last year, identified a different practical use for the content of the course. “I’ve used probability a lot more in everyday life,” he said, “especially since I got big into poker last year. It helped me out with that, with calculating odds and such.”

Michael took the course because he needed a university-level math credit in order to apply to the program that he wanted. “I do struggle at math a little bit,” he explained, “[and] it was a course that I heard I could get a decently relevant mark in.” For him, it was harder than he had heard from others, “but if you do your work, […] you’ll easily be able to get over an 80.”

Funk had a similar view of the difficulty level of the course. “It’s a U-level math course, so it’s moderately difficult,” she said, “but of the three 4U math courses it tends to be considered one of the less difficult ones.” It has, as Ms. Funk called it, a “moderate level of difficulty”.

“I think some take it to be a bird course for […] its accessibility,” Lucas explained, “but I see that more as a testament to how well it’s taught.” The content is so important in his eyes, though, that he would recommend it “not just for people who want an easy math credit, but for people who are looking for a better way to help understand their place in the world.”

Data Management is a useful tool to have under anyone’s belt, and the course at WCI seems to be quite effective at teaching it. If you are worried that statistics this political season may mislead you, or even if you are not, MDM 4UI is a course worth considering.