Mark Smith

The importance of conducting proper samples is highlighted within each of the readings. When informing the public on gathered information there are many steps that those conducting the polls or surveys need to maintain to make sure their information is accurate. In all three of the reading we see that the mistakes that are made within polling can ultimately lead to a bias or false reporting of our population. This in many ways comes down to the sampling error, but also in the overall understanding or lack there of when dealing with polls and our population.

Sheldon Gawiser and Evans Witt talk about the importance of the sampling error, when sample size is too small, one cannot get a just survey. Cluster samples are one of the highlighted and more effective sampling methods when trying to get an accurate account for all groups of the population. Cluster samples are used when randomly selecting certain numbers of people after putting the population into certain groups first. This is just one of the many sampling methods that Sheldon Gawiser and Evans Witt talk about in their section. Each method holds a different tasking and importance when dealing with public opinion.

I think in most cases the authors agree that polling should be a random process to insure lack of bias. It is important to maintain a certain idea of polling that shows the population for what it is. While looking at a couple polls on CNN.com it seems that it is almost impossible to have a really accurate poll. I mean who is taking these polls. There was one poll in the archives about illegal immigrating, and I was thinking that today a lot of the U.S has a large Hispanic population, it probably isn’t that one sided, if the polls were done with that in mind. I think it is hard to take polls because of the changing demographics.

In each reading we see that there is also another factor to consider when taking a sample, the sample size must be big enough to truly reflect the population. If the sample size is too small your sampling error margin goes up. To fully get a grasp on the population the sample size for any good survey or poll needs to be large enough to decrease your error.

Asher’s key points are numerous, as he talks about many different kinds problems that lead towards bias and inaccurate polling. Asher points out and talks about wording and an idea of leading words that in most ways demand a certain answer, this can be done with double negatives. Using this strategy a survey can lead its taker towards answer by to put it simple trickery. There was also an interesting idea of questions that have two thoughts presented, this way people that feel a certain way towards one idea might feel a different way towards another, giving the survey a bias direction.

The last reading from Erikson and Tedin stress the importance of the simple random sample. This is when “each unit of the population has exactly the same chance of being drawn as any other unit” (Erikson and Tedin). This is important when trying to prevent bias, if you are sitting in front of the science building asking people what they think of science when they come out it will probably be different than going to the athletic buildings. So there needs to be randomness to what or who you are drawing your data from.

In the end I think that each author would agree that there is a lot to know and understand when attempting to conduct a proper poll. It is important to understand that polls and surveys need to be accurate and lacking bias and error, because if they are not we cannot fully understand the importance samples have when looking at the entire population.

## Leave a Reply