# Learn: A Tutorial on Running and Interpreting Frequencies and Bar Graphs

In the practice you just completed, you should have found the variable names for the variables in the General Social Survey about serving on a committee and as an officer: "SERVEGRP" and "LEADGRP." You should have also noticed that searching in the General Social Survey follows a similar logic to searching in the DDB data, even though the former contains a lot more variables.

You also learned that if you select the variable name in the codebook window, it shows more detailed information. This information is called a frequency distribution . At this point you should consult your favorite statistics book about frequencies and, in addition, familiarize yourself with an important companion concept: level of measurement .

Rather than relying upon the frequency already produced for you in the codebook, there is a preferable way to run frequencies yourself, which gives you more investigative power.

Let's look at an example for running a frequency on SERVEGRP or "Has R [respondent to the survey] served on group committees."

After you open the General Social Survey SDA dataset and select "Frequencies or crosstabulation (with chart)" and the "Start" button, the following screen will appear where you can type "SERVEGRP" next to "Row." In addition, SDA allows you to create special types of bar graphs to better visualize your results. For this example, under "CHART OPTIONS" select the "Bar Chart" option next to "Type of charts."

Next select the Run the Table button at the bottom of the screen and the results will look like the following:

Let's interpet these results.

The first thing you should notice is that a code of '1' means that an individual has served on group committees and '2' means that an individual has not. This coding reveals that the variable "SERVEGRP" is measured at the nominal level of measurement .

Most important, the frequency reveals that a healthy 60.7 percent of the sample surveyed has been engaged in civic life in this manner. Moreover, in the bar graph you can see how large 60.7 percent is compared to those who did not serve on a group committee. This simple visual representation (one can immediately see that there are more who answered "yes" than "no") is often what makes graphs, charts, and tables a powerful way to communicate quantitative information. Visual representations of quantitative information often are underutilized (see Edward Tufte , a scholar who writes about this). In Bowling Alone, because Putnam writes for a more general audience, he pays particular attention to this, and hence presents a lot of the quantitative evidence visually.

On page 59 of Bowling Alone, Putnam rounds the figure 60.7 percent up to 61 percent. Also notice that this is just for the year 1987. If you read on to page 60, Putnam explains what happens in the next decade in America to this picture of what he calls "active participation," and it is represented visually in Figure 10.

Before we move on, there is still more to learn from this frequency. How many individuals were surveyed? An exact 1,174 individuals answered this question on the survey. Why so small? Again, if you delve into the codebook a bit further, you will learn that the variable "SERVEGRP" only occurs in the year 1987, when the GSS conducted a special investigation into civic engagement. This is why Putnam has to use other datasets and variables to produce Figure 10 on page 60. But for now, congratulations are in order. You have been able to replicate some of Putnam's work and apply knowledge about frequencies and level of measurement to your investigation. Now return to Exercise 1 to practice replicating Putnam's reported frequency on serving as an officer in a group in 1987.