If you need to create a datafile of your survey responses, the best approach may be determined by the analytical tool you plan to use. If you're going to do simple spreadsheet sorts and counts, you might want to enter the data right into a spreadsheet. Some databases and statistical software packages have data entry functions or add-ons that can be used.
Database packages often have components to make data entry easier, and some have simple reporting functions as well (eg Filemaker or MS Access). Or the database can then be exported for analysis using a different tool.
Data can also be entered into a text file (using Notepad on a PC or TextEdit or TextWrangler on a Mac, for example), that can be read by a statistical package.
One of the most flexible options is to enter the data into a spreadsheet, where each column represents a question (or question option, in the case of a "check all that apply" item), and each row represents a respondent. This format can be imported or read directly by many other packages, or the file can be saved as a .txt file.
What tool you use to analyze your survey data depends on the purpose of the survey and the kinds of analyses you wish to perform. Simple descriptive summaries can be accomplished with a number of packages, but more sophisticated statistics will require a statistical software package.
(Note that, unless otherwise stated, the links below take you to the vendors' websites for more details about the packages. But all these packages are supported to some degree by ITS, so you wouldn't want to purchase anything before checking into availability and access in the "Supported Software" section of the Information Technology Services (ITS) Website.)
Swarthmore provides access to SPSS and STATA through the key access server (see the ITS Website). Both of these are excellent interactive data analysis programs. Princeton University has an online STATA tutorial that you may find helpful.
Data Desk is also an interactive program, but it doesn't assume any specialized training in statistics. It's very good for exploring data as it comes in, and is particularly adept at graphing and manipulating graphs on the fly for visual data discovery.
JMP is another package which is quite good for analysis and is perhaps easier to learn than SPSS, which relies on your knowing statistical terminology. But an advantage of SPSS, or Stata, is that they can save analytical steps into procedure files that can then be called up again as needed. If there are programming tasks for analyses that you expect to repeat over and over again, this could be valuable.
Finally ITS also has KaleidaGraph, a general purpose data graphing tool. (Many of the other packages also support graphing, including Excel.)
For the average occasional survey analyst, JMP or SPSS provides a great combination of user-friendly menus and statistical strength. Visit ITS for more details about supported software. Also, Doug Willen (firstname.lastname@example.org) in ITS, can take your questions about statistical packages.