You can do this if you are very careful. First, create a study with the same:
- attributes (order does not matter)
- levels (order should be increasing or decreasing if it is a continuous attribute. If it was $15, $10, $5 in the original study, it is okay to put in $5, $10, $15 as long as you code the levels correctly, but you wouldn't want to enter them as $10, $5, $15.).
- number of choice tasks (must be the same as the original)
- number of concepts per task (must be the same as the original)
- number of versions (can be different if you want to code each respondent's design as a separate version, even if they're the same)
Next, create a design file based on the paper and pencil surveys, and import it into your new CBC survey.
Finally, import the paper and pencil data as you normally would.
This, of course, assumes that the design file is indeed lost. Before I went too far in manually recoding the entire thing, I'd check the study backup folders to make sure there isn't a copy there. If you have your .ssi file, you can also try regenerating your design, and then compare it to your paper surveys to see if it matches up. (If the design hasn't changed since you created the paper and pencil files, it is very likely that regenerating will produce the exact same design.)
Please call us if you have ANY concerns about how to do this, as any mis-steps can lead to invalid data.