The purpose of this chapter is to provide you with enough information to begin to read, understand, and evaluate published research in the social sciences. It is not a comprehensive introduction to statistics; for more information, see the ‘Further reading’ suggested below.

The first part of the chapter reviews some common methods for summarizing data, both numerically and graphically. The second part gives an overview of significance testing: why it's important, how it's usually reported, and how to interpret it.With the background provided in these two sections, you should be able to read a typical scholarly paper reporting the results of social science research and understand the gist of what's going on, even if you cannot follow all the details.

The last part of this chapter describes some of the major issues in conducting social science research. This section will expose you to some of the most important questions you can ask about how an experiment was designed and carried out; learning to ask those questions is the first step towards being able to evaluate the results of a particular study.

The examples in this chapter are taken from a study of the effects of intoxication on speech. In that experiment, eight subjects recorded a list of several dozen words, once while sober and once while moderately drunk; see Kaplan (2010) for a more detailed description of the methods involved. I've chosen these examples for convenience: the experiment generated a large amount of data that lends itself to many different types of analysis. The results presented here are meant to illustrate statistical points; they aren't novel (or even representative) findings about the general effect of alcohol on speech, nor are they necessarily the best way to analyze the data. For an overview of the literature on intoxicated speech, see Chin and Pisoni (1997).

Descriptive statistics

The purpose of descriptive statistics is to summarize datasets in a concise way. This section reviews some common descriptive techniques, both numeric and graphical.

Some common statistics for summarizing data

Perhaps the most common statistic for summarizing a large set of numbers is the familiar mean (or average), which is a single number intended to reflect a typical value for the set. Mean is often abbreviated with the capital letter M.