Sampling: How do I Choose a Sample? (Part 1)
Before you begin your research, you’ll want to think about how you are going to get the sample you need. How the sample is chosen is very important because when it is done correctly, the sample will accurately represent the population from which it came.
It might make more sense to start with the wrong way to get a sample. One of my local television news programs frequently presents a topic to “the citizens of Lubbock” and then asks for our opinion, usually whether we agree or disagree. Viewers are invited to log on to the station’s website to register their opinion, and the news anchor reports the results on the next day’s news. This method almost guarantees the sample of opinions will not represent the people of my city (that would be Lubbock). There are several problems with this approach, but the main problem is that for a sample to represent a population, every person must have a chance to be heard. Every person doesn’t have to be heard, they just need to have the same chance. If they don’t, certain types of people will likely be missed and won’t be included in the sample. I think its safe to assume that since there are several news programs available, everyone in Lubbock doesn’t listen to that particular news program. Even if they did, there would surely be some that didn’t tune in on that day, or maybe they don’t have access to a computer. To give credit to the news program in question, however, they do provide the disclaimer that their methods are not scientific. So, they are forgiven J
But this does make the point that choosing a sample should be done correctly in order to get results from your study that you can trust. In the previous blog I mentioned a fictitious dentist conducting research on his patients concerning whether reminder calls reduced the number of missed appointments. Instead of selecting his sample from patients making appointments in March, July, and November, why not just take all the patients making appointments in the first month of the year? The reason is that people making appointments in January may be different from those who make appointments throughout the rest of the year. What if several of these patients are new-year’s-resolution-makers who are merely fulfilling their resolution? Wouldn’t we expect these people to be more likely to keep their appointments even without a reminder call? Of course, and this may make it appear that reminder calls were effective when these people would have shown up anyway.
There are many considerations for choosing a sample, but the main one is that we need to make sure everyone in the population we are drawing from has an equal chance of being chosen for the sample. This ensures that we select at least one representative for each of the different kinds of people that make up any population. In the next blog I will discuss a few of the more popular methods for sampling.