Coding sheet

Once upon a time, I was providing statistics consulting to a young PhD candidate who was working on his dissertation. One of his analyses involved comparing men and women on some outcome, and his dataset included a variable named "gender" with the response options "1" and "2." When I asked him if men were coded as "1" or "2," he didn't know. He'd entered the data years ago when he was working as a research assistant, and he couldn't remember. To make it worse, he had thrown away all the original surveys that he'd used to gather the data. So there was literally no way to figure out if "1" meant "man" or "woman." He couldn't do the gender analyses, and now he's managing a Best Buy in rural Missouri.

This is why the first thing we do before we conduct any statistical analyses is to create a coding sheet. A coding sheet is a spreadsheet that has information about your variables, including a name for each variable, a short description of the variable, the type of scale that was used to measure the variable, and the coding you used for the variable. This ensures you will never forget if you coded "male" as 1 or 2, even if you put your dataset away for years. It also ensures that if you ever need to share your data, whoever you share it with will have a resource to know what your variables mean and how they were coded.

For your next activity, please take a look at the coding sheet that I created to describe the variables from The Progressive Group survey, watch the video describing how I created the coding sheet, and then use Excel to create a coding sheet of your own describing the variables from the Commit Corporation survey. It will probably be useful to have ready access to The Progressive Group survey and to the Commit Corporation survey.

Here is the coding sheet that I created for The Progressive Group survey responses

Here is a video explaining how I made the coding sheet