April 24, 2013

Information is not data

A wise professor once reminded me that information is not data. Stacks of videotaped interviews are information. In order to analyze it you must first turn it into data by viewing and coding it and recording the codes in some analyzable form. That process is not analysis: it is coding. Once the coding is done and the information has been turned into data, the analysis can begin.

My graduate education was so steeped in that wisdom that I can hardly understand people who don't make that distinction when they talk about research, analysis, information, and data. If I send someone a summary table and they thank me for the "data" I am confused. When I ask for data and I get a list of means, I get irritated.

And when someone gives me an Excel spreadsheet that contains organized fonts and colors and column widths and combined cells and multiple lines of header information and rows of summary statistics for subgroups (indicated by cell borders) and so on, I begin to get downright angry. That spreadsheet may well contain all the information I need in order to extract the data, but the author has spent what is no doubt an aesthetically enjoyable time obscuring that information as much as possible.

Data are square. Data do not have curly flourishes. Data are not pretty or elegant or understandable at a glance. You should spend as much time formatting your data as you do drawing on your windows. Oh, you don't draw on your windows? Why not? Because you want an unobstructed view of the world? Ah yes. So you do understand.

Information is nice, but data are useful.

Posted by rbrent at 02:46 PM | Comments (0)