Social data analysis enables you, as a researcher, to organize the facts you collect during your research. Your data may have come from a questionnaire survey, a set of interviews, or observations. They may be data that have been made available to you from some organization, national or international agency or other researchers. Whatever their source, social data can be daunting to put together in a way that makes sense to you and others.

This book is meant to help you in your initial attempts to analyze data. In doing so it will introduce you to ways that others have found useful in their attempts to organize data. You might think of it as like a recipe book, a resource that you can refer to as you prepare data for your own consumption and that of others. And, like a recipe book that teaches you to prepare simple dishes, you may find this one pretty exciting. Analyzing data in a revealing way is at least as rewarding, we’ve found, as it is to cook up a yummy cashew carrot paté or a steaming corn chowder. We’d like to share our pleasure with you.

Quantitative or Qualitative Data Analysis?

Our book is divided into two parts. One part focuses on what researchers call quantitative data analysis; the other, on qualitative data analysis. These two types of analysis are often complementary: the same project can employ both of them. But for now we’d like to look at the main distinction between the two. In general, quantitative data analysis focuses on variables and/or the relationships among variables. This analysis involves the statistical summary of such variables and those relationships. Roger recently completed a study, with two students, of the relationship between Americans’ gender and their party affiliation (Petit, Mellor and Clark, 2020). We were interested in the relationship between two variables: gender and party affiliation. Women are more likely to identify as Democrats in the United States than men. We found that we could largely explain the emergence and maintenance of this relationship since the 1970s in terms of three other variables: the increased participation in paid labor by women, the decreasing likelihood that both men and women are married, and the declining participation of Americans in labor unions. To examine the relationship among these five variables (gender, political affiliation, labor force participation, marital status and attachment to labor unions) we relied almost exclusively on quantitative, or statistical, analysis.

Qualitative data analysis, on the other hand, focuses on the interpretation of action or the representation of meaning. Roger did another study with a student in which we watched YouTube recordings of Trump and Clinton rallies during the 2016 presidential campaign (Fernandez and Clark, 2019). A careful examination of these rallies led us to the conclusion that Trump rallies typically looked more like quasi-religious events—with participants displaying quasi-sacred objects (like Make America Great Again caps), participating in quasi-religious rituals (like shouting rhythmically and in unison, “Lock Her Up”), and cheering quasi-religious beliefs (like how valuable it would be to slow immigration)—than Clinton rallies did. In this study, we were focused on both interpreting the actions of rally participants and on trying to represent what they meant by those actions.

In practice, researchers often employ both quantitative and qualitative data analyses in the same study. For instance, Roger recently did another project with students^[1] (Gauthier et al. 2020), one of whose goals was to discern the decades (one variable) since the 1930s in which children’s picture books were most likely to depict characters in gender stereotyped ways (another variable). Put this way, the study looks like one that required quantitative data analysis, because it was examining the relationship between two variables—the time in which books were created and the degree to which they depicted characters in gender stereotyped ways. But to discern whether an individual book used gender stereotypes, we had to interpret the actions and thoughts of individual characters in terms of a number of characteristics we viewed as gender stereotyped. For instance, we had to decide whether a character was nurturing (a stereotypically feminine characteristic) and whether they seemed competitive (a stereotypically masculine characteristic). Such decisions are essentially qualitative in their nature.

Consequently, the distinction we’ve used to organize this text—quantitative vs. qualitative data analysis—is a little misleading. Researchers often employ both kinds of research in the same project. Still, it is conventional for teachers to teach quantitative and qualitative analyses as if they were distinct and who are we to defy convention? Thus, this text includes both chapters about quantitative data analyses and those about qualitative data analyses.