Kelly O’Neill

March 2, 2010

Reaction Paper

Sampling, an Important Tool in Polling

Sampling is one of the most important factors associated with polling and with analyzing poll results. A sample is a smaller group of people used to represent the entire population or at least give a good idea of what the population’s general behaviors and attitudes are. In “The Poll: Sampling Error” by Gawiser and Witt, sampling error is discussed and explained. The authors mention that no matter how good a survey is, there will always be some error – even if the survey is conducted perfectly and without any mistakes. There will always be error because not every single person in the population was surveyed; therefore, there is the likelihood that the poll results may not match the true opinions of the entire population exactly. Sampling error is defined as the “amount of chance variation expected in a series of samples in the population” (Gawiser & Witt). Sampling error is dependent on two criteria: the size of the sample and the size of the population. This means that no matter how well the researchers conduct their survey, the margin of error will vary from anywhere between .001% to 7% (5% is the most that most researchers will accept to still refer to a poll as accurate). As a journalist, it is very important to make sure that readers understand that this sampling error (among others) exists, and that they are not led to believe that the poll is 100% accurate or representative of the entire population.

In Austin and Pinkleton’s “Making Research Decisions: Sampling,” the authors went into a general discussion of what sampling is, and they defined and explained several of the terms associated with sampling. There are two main types of samples: probability samples and nonprobability samples. Probability samples are random (every person within the population has an equal chance of being chosen to participate in the survey/poll), while nonprobability samples are not random (some people have a better chance than others of being picked to participate). Obviously probability samples are more accurate and are a much better representation of the whole population, but they are also more expensive and time-consuming. The authors also discussed three misconceptions about sample sizes. The first is that bigger samples are better. They explain that if a sample is not representative of the group, then size is completely irrelevant. The second misconception is that as a rule of thumb, researchers should sample a fixed percentage of a population to produce an acceptable sample size. The final misconception says that researchers should base sample sizes on industry standards or “typical” sample sizes used in other research projects. Another facet of sampling discussed in this chapter was sample distribution. Sample distribution is a “grouping or arrangement of a characteristic that researchers measure for each sample member, and it reflects the frequency with which researchers assign sample characteristics to each point on a measurement scale” (Austin & Pinkleton). Sample distribution can be represented by a bell-shaped curve, which is often used by researchers. Standard deviation is the “standardized measurement of dispersion (or variation) around a mean” (Austin & Pinkleton). Basically, standard deviation measures error. The number most often used as the norm for standard deviation is 1.96 which corresponds to a 95% confidence level (only 5% sampling error). This chapter greatly helped in explaining the basics of sampling and how it can be used effectively by researchers.

I personally think that sampling is a great way to get a feel for public opinion. I think that it is an easy way to represent the population as a whole and show the public different trends that are going on in their town, state, country or even in the world. One part of sampling that I would like to focus on is how journalists should use polls and survey results in their stories. I think that it’s very important for journalists to make sure that their readers/viewers fully understand what they are looking at when they look at poll results. This is not only includes what the subject of the poll is or what question was asked, but also what the sampling error is, who was surveyed, etc. Gallup Daily Polls’ website does a great job of explaining all of its polls (and there are a lot of them) to its audience. For example, a poll on gallup.com that covers the issue of unemployment in the United States comes with what the margin of error is (+/- 0.7%), and also what the researchers consider “employed” and “unemployed.” It also explains how the survey was conducted. Because of this information, the poll becomes much more meaningful to the general public, and it also seems much more credible. Audience members can know that Gallup isn’t trying to deceive them in any way. This is what journalists need to be sure to do when reporting a poll story.

## Leave a Reply