Opinion Polling

1. Opinion Polling The who and how.

2. Who does polling? News organizations like CNN, Fox News, ABC, and NBC. Polling organizations like Rasmussen, Gallup, and Survey USA. These organizations use slightly different techniques when polling, but their methods are fairly similar.

3. Samples Pollsters take a sample of the greater population. There are various kinds of samples, four broad categories exist in two exclusive sections. They are: Biased and unbiased samples. Random and nonrandom samples.

4. Nonrandom or Random? Random samples are those selected at random from the population as a whole. Nonrandom samples are chosen with some parts of the sample manipulated. Nonrandom samples are not bad samples! Sometimes polls would not be accurate from a random sample. Political polls are almost always nonrandom due to “likely voter” selection.

5. Biased or Unbiased Samples Biased samples are bad samples! Biased samples are chosen in a way as to make the poll inaccurate. “Convenience samples” like internet polls are biased samples, not because they support an argument but because the method will not generate a representation of the population. Unbiased samples are good samples! Unbiased samples are those taken in such a way that they accurately represent the population as a whole.

6. Likely voter selection Instead of a random sample participants in campaign polls are determined to be likely voters before inclusion. The method of determining likely voters is simple among most news organizations, they ask during the telephone survey.

7. Margin of Error All polls have a margin of error, this is determined by the sample size used in the calculations. The larger the sample size becomes the smaller the margin of error is. The costs of the poll itself also increase though. News organizations balance costs with low margin of errors when conduction polls.

8. Poll Manipulation Some polls are manipulated, most of these by using biased sampling methods. Internet polls or weighting votes based on party are common ways of achieving biased samples.

9. Physically conducting the poll. Most polling organizations and news groups use a very similar polling method. Using random dialing from a phone book list, they find and weight a sample based on likelihood of voting. They determine this weight and whatever other information through their survey. The numbers are then crunched to come up with the final values.

10. Question Is the class a good sample? If it is, then of what population? If it isn’t, why not?