Saturday, February 21, 2009

Week 7: Surveys

Surveys are a commonly used research technique and are relatively inexpensive to implement. If used correctly, surveys can be generalizable and can allow researchers to gather information about a very large population. To do so, the sample chosen from a population must be randomized. Even when choosing a random sample, a large enough sample size must be found for the results to be reliable. According to Lauer and Asher, "The precision of information is directly related to the sample size." There are several methods to approximate a random sample group including systematic (e.g. every 10th subject is chosen), quota (e.g. including 10% of a certain group in your sample, if the entire population has 10% of that same group), stratified sampling (e.g. focusing your sample on one specific group, but still studying the larger population in which that group resides), and cluster sampling (e.g. choosing an entire classroom to study a population of the whole school rather than randomly selecting students across the school).

Another important part of surveys that researchers must keep in mind is to make sure that they have specific definitions for their sample, units, population, and their measurement methods. If possible, it may be a good idea to use a pre-existing, proven measurement technique. Lauer and Asher also warn that a low response rate can make a study's results unreliable and that if a new measurement method (a new survey) does need to be created, it should be tested on a representative group before hand, so that it can be adjusted for the best results in the real survey.

This chapter is especially useful to the research group that I am currently in for this class. As of right now, our plan is to survey undergraduate students at Clemson to get their opinions on possible methods of communicating to students that they have an upcoming tuition bill. Before reading this chapter, I did not think of the fact that a low response rate would lead to unreliable results, but just thought that it was an unavoidable, but unimportant part of the surveying process. Additionally, I did not think it was necessary to find a sample group from our sample group to do a test run of our initial survey, but upon reading this, I can certainly see the importance of such an activity. We would hate to either spend hours collecting survey answers or even worse, spend our one chance at getting the university to send out a survey on our behalf, only to find out that our questions were not properly worded. These issues pose even more challenges to our project in class. Frankly, I'm a little worried, but this chapter was certainly necessary for our group to read before conducting our research.

No comments:

Post a Comment