Survey Software Data Management II
This blog continues the discussion of data management started in the previous blog.
Another ability some survey software users need is to perform mathematical transformations on their data. For example, they may want to add up the answers to several questions to produce a total. Or they may wish to find the difference between two numbers or between two dates. A sophisticated program should let you perform all these calculations, as well as multiply and divide variables either by each other or by constants.
The most flexible of these kinds of transformations lets you define a new variable based on the combinations of answers to other variables. For example, you may have a gender question and an age question and would find it useful to have a combined gender/age variable. Or you may choose to compare women in certain age groups who have purchased a particular product with women in the same age groups who purchased a different product. And you may wish to compare them with women who have not purchased any product.
A related ability is to be able to create lists of people who gave combinations of answers. For example, you may notice on a table that some people seem to have given inconsistent answers. You can ask a capable program to give you a list of the ID numbers of all the people who have given a a particular combination of answers. That way you can investigate all the answers they gave and potentially decide to change one or more answers.
This brings us to the next ability – the ability to edit individual data records. Some programs might let you pull up an individual record and go through their interview again, changing anything you believe should be changed. Other programs can present all the answers in the form of a spreadsheet. The most sophisticated survey software offers both options.
Another type of data management is the ability to merge two or more data files. You should be able to merge in two different ways. The first is to add more records to a data file. Perhaps you had different people create two or more data files for the same project. In this case, the data files would have different people answering the same questions. You might want to merge these files into one for easier analysis and backup. The other way to merge data files is to add new information about the same people into an earlier file. For example a project may require you to recontact the same people multiple times to see how their answers change over time or how they respond to new questions about related topics. In this case the program must be able to match up the answers from people in one way with the answers given by those same people in a later way. The most capable programs let you do that, even though you may not be able to contact all the people in the first wave in one or more later waves.