 In this video, we'll use SOFA to analyze categorical data, using the item in the survey that asked people to pick their favorite color. First, let's find out if there's a significant difference in color choices by gender. Since this is categorical data, you want to use the CHI square test. Choose the correct table, and then the variables color and gender. Then show the results. Even though the graph shows a large difference in the proportion of males and females who like purple, overall there's no significant difference in color preference based on gender. We notice because the probability value is greater than .05. By the way, the color names are shortened to five letters, because that's all the survey software allowed us to use for the answer code. The next question is whether there's a significant difference in color choices overall, or are all the colors equally preferred? If you look at the totals in the table, you see a small number of people who like orange and yellow. Does that make the overall difference significant? To do this, you again need a CHI squared test, but it's a one variable test, and SOFA doesn't do these, at least not yet. Instead, go to this URL, SOFAstatistics.com slash CHIsquared.php, and enter the totals for each color along the top row, and then click Calculate. In this case, there really is a preference for colors, and it's significantly non-uniform, and we know this because the probability value is .001.