How to Enter Data from Paper Surveys
You’ve done it! You’ve collected data from 500 paper survey respondents and it’s all recorded onto paper forms. It’s time to start doing the analysis, but before you can create your report you need to get the data into a format that you can use in analysis software.
Depending on your budget, perhaps the form you have is machine readable, and in that case you probably have software to take care of the scanning already. In most cases, however, you may find that the data will need to be entered manually. You can do this in a simple text editor or in spreadsheet software, for instance. Many prefer Excel and I’ll use that as the example here.
Assume that your survey has 5 simple questions, where you ask the person’s name, gender, age, favorite sodas from a list of five options, and which part of town they were in when the survey was conducted. In Excel, you would set up column headings as shown here:
But something’s missing! You should be able to go back to check that the data has been accurately entered for each survey and you need an identifying number that appears both in your Excel data file, and on your paper questionnaire. Add in a column for SurveyID, so that you get:
Entering the Data
Now it’s time to enter the data. The survey ID can be sequential, as long as the numbers are unique for each survey. The name of the person (or any type of text data) should be entered exactly as it is on the paper form, even if there are spelling errors. It’s important to resist the temptation to correct people’s spelling, especially if they’ve filled out the form themselves. For age, type in the age they’ve provided. The gender data can be typed in as a 1 for “Male” and a 2 for “Female” if you prefer to use numeric coding. However, by typing in the text label instead, you may save yourself some time to enter these labels later in your analysis software. Remember that Excel auto-fills the whole word if you just type the first letter, so it won’t take any extra time to enter M or F instead of 1 or 2. The same idea goes for the location question.
The trickiest question here is the multiple-option question about soda types that the survey respondents like. On your form, maybe they’ve just ticked two things and left the rest blank. That means that they’ve answered the question and so you should enter a 1 (or a label, such as “Selected”) into the columns that correspond to the soda that they like. All other response options should have a 0 if the respondent has not selected them. If a respondent doesn’t select any, then they haven’t answered the question, and the whole row should be left blank. In other words, a row should either be entirely blank, or have a mixture of 1’s (for “Selected”), and 0 (for “Not Selected”).
When you’re done, your file might look something like this:
Don’t forget to save your file as you go, and when you’re done you can import this into your favorite software for analysis.