Survey data analysis
The ease of basic data display and analysis is one of the reason I love online surveys. The reason for this ease is what’s behind the survey. The bottom line: survey answers go directly into a database. You must still, of course, design items that will produce data answering your critical questions – that kind of design task is just as difficult in an online or other computer-based survey as it is for hard copy surveys. But, if you design your questions well, you will be able to use the survey application to create data displays that make sense and allow visual analysis of data, as well as conducting basic descriptive statistical analysis directly in most survey applications.
Downloading item responses
As with other things we’ve discussed, display capabilities vary among the online survey applications. I will discuss SurveyMonkey as an example. Most subscription applications will have at least the same capabilities; more expensive applications will provide even more options. I will use SurveyMonkey as an example of what’s available, beginning with the options for viewing and downloading responses; those options are listed in the screen print below:
(Click image for a larger view.)
In my own use of this application, I have often found it useful just to browse responses. The application allows the user to browse across all respondents on a single item, or to browse responses to all items for a single respondent. This kind of dip-sticking provides some interesting insights while responses are still coming in without compromising the database.
SurveyMonkey provides an option for filtering responses, as indicated in the screen print below:
(Click image for a larger view.)
As stated, responses may be filtered by response so that, for instance, only people who answered in the positive to a particular item will be in the sample. Responses may also be filtered by the properties of the response; this option is nice because it allows limiting an analysis to responses between certain dates or other delimiters. As with analyses applied to an Excel or Access database, the analysis does not in any way alter the basic data stored in the data base; if what you try doesn’t provide the picture you need, simply discard it.
The crosstab option allows another way of comparing responses of a limited group of respondents against the total pool. The following screen print from SurveyMonkey documentation provides a clear and succinct comparison of filters and crosstab possibilities:
(Click image for a larger view.)
One option is to export responses to each item individually. SurveyMonkey provides several export formats, as is indicated in this download screen:
(Click image for a larger view.)
The general options allow a general summary report (narrative) or a download of all data collected to date into a format that can be imported into a database or spreadsheet for further analysis. As you can see, several of these options provide data in a database format; these options are particularly useful if you will be subjecting the data to more rigorous statistical analyses not available in the online application. In addition, the responses may be downloaded as an HTML object that may easily be posted to a website. A nice option is the ability to download the open-ended questions, in whatever format you’ve chosen.
Next up: Graphic data displays from SurveyMonkey