Hi, I’m Ben Owens. I’m a math teacher, and in this video I’ll explain how statistics apply to political polling. During political campaigns, the media, the voters and the political parties want to know how the public at large thinks about certain candidates or policies. One way to find out would be to knock on every door in the country, asking every citizen their opinion. Of course, that wouldn’t be practical. Luckily, due to the laws of probability and statistics, we can determine voter preference by using something we call a statistical sample. When conducting statistical samples, Polling companies ask a relatively small number of people questions like: Who is your favorite candidate? or Who are you most likely to vote for? And, the opinions of this smaller group of people can be used to represent the population as a whole. Using a small sample to represent the opinions of a larger population is called Inferential Statistics. So you might be thinking, “How can we be sure that a sample will be representative of the entire population?” To be honest, you can never be 100% sure, but using probability and statistics allows us to make a prediction and mathematically measure how confident we are in our prediction. Of course, there are a lot of factors that need to be taken into consideration to ensure the political poll is accurate. First, care must be taken to make sure the sample is representative of the population. To achieve this, some sort of random selection must be used. There are different methods of random selection (simple random samples, stratified sampling, cluster sampling, etc). But, if the sample is not chosen at random, then the laws of probability and statistics do not apply! Another factor to consider is the timing of the poll. When news breaks, or an issue pops up, voters can be more sensitive. Thirdly, you also need to know what methods were used to conduct the poll. Was the poll done online by an opinionated political blog? Or was it done by an independent polling firm, calling random phone numbers? And lastly, the size of the sample should be taken in consideration also. A small random sample of, say, 20 voters might not represent the larger population. A larger sample, when the poll conducted well, is always preferable. For more on statistics, vist education.about.com. Thanks for watching!

#### Discuss

0 comments characters remainingSubmit