Descriptive and Inferential Statistics: What’s the Difference? (Part 2)

Descriptive statistics should be included in every research study because they help give an accurate picture of the sample, the measurements used in the study, and results.  But most research studies go beyond descriptives and also use inferential statistics.

Inferential statistics are numerical values, like descriptive statistics, but they are derived from statistical analysis tests.  Statistical tests use data gathered from a sample to mathematically determine if the sample is significantly different from the population from which it came.  They are called “inferential” because we use what we learn from the sample to infer what is likely happening in the population.  For example, if a dentist wants to know whether making reminder calls to patients the day before their appointments would decrease the number of missed appointments, he could take a representative sample of patients such as a selection of appointments made in March, July, and November of a given year and he could have his secretary give each one a reminder call.  Then, he could use a statistical analysis test to compare the sample’s missed-appointment rates to that of all of his patients. The information gathered from this sample would allow him to infer that the effectiveness of reminder calls for the sample would be the same for all of his patients.

Why don’t we just measure the whole population?  There can be several reasons, depending on the nature of the study, but the main reasons are because it would be too costly and it would take too long.  Another reason is that in many cases, we don’t have access to an entire population.  Gathering data can be expensive, and most answers we gain from research are needed as soon as possible, so using a sample to represent a population solves these problems.

Inferential statistics are what allow us to use sample information to generalize, or infer, what would happen in a population if the same treatment were applied.  This only works, though, if we choose our sample in the right way.  Next time, I’ll talk about how to choose a sample that will represent the population we want to study.

Comments are closed