Descriptive and Inferential Statistics: What’s the Difference? (Part 1)
When we hear the word “statistics”, most of us think of things like the average temperature for a given month or the average rainfall in a year. Here in West Texas we’ve only had about a half inch of rain this year which puts us about 2 inches behind (gasp!). I wouldn’t know this if it weren’t for the statistics reported by our weather guy every day. Or we might hear on the news that over 50% of Americans are on prescription drugs (can you believe it??). Information like this, delivered to us in numerical form, describes everything from the weather to the state of Americans’ medicine cabinets. Hence, they are known as descriptive statistics, or simply descriptives, and they give us a point of reference for the information we hear in our everyday lives.
Descriptive statistics are very important in research too. We use descriptives to describe our sample in ways such as the average age of the participants, the ethnicity breakdown, sometimes the average income, and other descriptives that are relevant to our study. This is important so that readers get a good picture of who the sample was, allowing them to put the results of our research in the appropriate context. For example, suppose a research study indicates that children who are presented with “green trees” on their plates instead of broccoli tend to eat more of the vegetable. We might think to ourselves, “this will never work on my 12-year-old”, and that may be true. It would never have worked on my kids. But if we know that the sample consisted of 4-year-olds, then the results make more sense. Descriptives are also important for replication. The results of one research study are never enough and should not be relied on for information. But when research is replicated, or conducted several times by different researchers, and the results tend to be about the same, that is when we can trust what the results tell us. Researchers who replicate the work of others depend on an accurate description of the original sample so that they may choose as similar a sample as possible. Otherwise, the results will surely be different and will tell us nothing about the reliability of the original work.
We also use descriptives to describe the instruments used in our study. The average before and after depression scores in a research study measuring the effects of counseling on depression in elderly people is very important to know. Knowing their average scores before and after helps the reader judge the practical effects of the study. Descriptive statistics are important in several ways, but most research studies go on to include inferential statistics as well. I will talk about those in the next blog, so stay tuned!