Before any data analysis can begin, you have to have some data to analyze. Where do your data come from, and how do you go about obtaining the data you need to test your hypotheses and answer your research questions? Data can take many forms, it can come from many diﬀerent sources, and there are various approaches to ﬁnding the data you want. As discussed in Chapters 1 and 2, data are the outcome of the measurement or categorization process. But before you can measure, you have to have someone or something to measure. In other words, you need to have recruited participants for your study or otherwise obtained the research units to be measured on the variables relevant to your study. In this chapter, I give a broad overview of the various approaches to recruiting or otherwise obtaining the research units that provide data to the researcher.
When we conduct research, we are often interested in some kind of inference. In statistics, the most common conceptualization of inference is population inferencethe practice of making a statistical statement about a collection of objects from a subset of that collection. For example, if 60% of 200 people you ask approve how the president is doing his or her job, you might make the claim that around 60% of all people (rather than just those 200 you asked) approve. In so doing, your claim that 60% of all people approve is a population inference. To make such an inference, you must be very careful in how you go about recruiting or ﬁnding research units to be measured (e.g., people who respond to a question or participate in the research in some fashion) because most statistical methods make some kind of assumption about how the units were obtained in order to apply those methods to making population inferences.