A common persuasive tool for speeches is to use statistics. Here’s a quick overview of statistics and some helpful thoughts on how you can do the math, making your speech more powerful.
Statistics is the practice of using sample data to figure out population data. The “population” is the entire group of people or objects that you want to figure out something about. The “sample” is a smaller sub-group of the population that you use to gather data for convenience’s sake.
What does that mean?
Here’s an example. If I wanted to find out what every person in Seattle thinks about a certain issue, that’d be a difficult task. How do you survey thousands of people without missing any? Instead, I might survey a few neighborhoods in that area and extrapolate based on their answers what everyone might think. This smaller group is the sample, the entire group is the population.
Two common myths about statistics:
1. Sample size is the biggest factor in the accuracy of statistics.
This may seem intuitive, but it’s not quite true. In fact, the most important detail in a statistic is how the sample was collected, not how big the sample is. The reason is complicated, but numerous studies have found it to be true. As long as the sample is greater than 30, it will predict the population data fairly accurately.
However, if the sample data is collected from a group that isn’t representative of the population, there’s your problem. If I wanted to see what percent of speakers have won awards in competitions, surveying Potent Speaking readers would be inaccurate because the type of people who read this website are more likely to win than other people.
2. Statistics can be trusted.
Especially when it comes to opinion surveys.
Wikipedia’s article on misuse of statistics reports nearly 12 different ways to misuse statistics.
It’s important to look for who commissioned the data and who collected the data. Who paid for the study? The bias of these people will almost certainly come through in the results.
An example of a commonly used statistic that is false is the claim that “97% of scientists believe that global warming is happening and is caused primarily by humans.” This statistic was gathered from a survey of over 3,000 scientists—but only 79 were considered in this statistic. They were specifically chosen by the guy who came up with this statistic. As plain as this fraud is, the statistic has been quoted by government officials including President Obama himself. (I’m sure they were unaware of the fraud, but it shows that just because someone credible said it, doesn’t mean it’s right).
The way the question is asked is extremely important. Consider the difference between these two: “Do you believe we should spend $1,000,000 to fix our roads?” vs. “Do you believe our roads should be fixed?” One question actually includes the downside to the proposition, and is much more likely to get a “no” answer.
The way the data is interpreted is important. The New York Times gave four good pollsters the same poll data, and they all came up with different results. The reason for this is that many surveys include fill-in answers that must be interpreted, or other such subjective data. This is one of the most common reasons for statistics that are rather bogus. Remember the global warming statistic issue? It was caused by subjective human tampering with the data.
The person collecting the data can easily skew results using a variety of hidden tricks.
1. If you want to discredit someone’s statistics, look into how the sample was collected, not necessarily the size of the sample.
2. If the sample size is ridiculously low compared to the population being evaluated (eg. 100 people to determine the entire United States’ opinion), then it’s worth attacking that. A big sample size does help to reduce bias, because it’s harder to choose 4,000 biased people than 100.
3. For polls, make sure the survey questions were fair and didn’t make certain types of answers more likely by their very nature.
Don’t miss future posts like this one, subscribe to my email list below!