In order to begin this discussion, we need to establish terms. The population is the group of individuals we would like to aquire information about. We go about this using samples, which is the part of the population we examine in order to gain information about the whole. The design is simply the method which one goes about gathering samples. There are many different designs - some situations require one, and other situations require others. In some cases, a one design holds more bias than other, and so the other is chosen. A design is bias if it systematically favors certian results. For example, if you wanted to see who people were voting for president, and you gained your sample by selecting people who had a telephone number. This design is biased because it systematically removes the poor from the sample. As you can see, one needs to be careful about their design choice in order to prevent biases.
The fundamental sample is the simple random sample (SRS). This design gives every possible sample an equal chance of being selected. It is put into action by giving everyone in the population a number and then using a random number generator or chart to choose the sample. A branch off of the simple random sample is the stratified random sample, which divides the population into groups, and then randomly selects people from these groups. This type of sample is useful if you want to make sure to get a sample of people from a certain group. For example, if you were testing a new form of medicine and you wanted to make sure it works on both male and females, you could divide them into two groups and then select them with a simple random sample to ensure that you would get the same amount of men as women.
Choosing the proper design is crucial in avoiding bias, as stated earlier. There are several forms of bias, namely nonresponse, reponse bias, and misleading questions. These forms of biases can pop up in the surveys and the design and can ruin your data! That is why we are learning this now!
