Is there a better way for inputing data manually?
At 04:40 02/12/2006, Duncan Murdoch wrote:
On 12/1/2006 11:00 PM, Wei-Wei Guo wrote:
Dear All, I have worked with R for some time. It's a great tool for data analysis. But it's too hard to inputing raw data manually with R (I don't mean importing data. R is good at importing data). Maybe it's not a focused topic in this list, but I don't know other place where I can ask the question. How do you do when inputing data from a paper material, such as questionnaire, or what do you use ?
I would not use R for this. Depending on how many questionnaires I had, from small number to large, I would use: 1. A text file. 2. A spreadsheet, like Excel, or the OpenOffice one, or the R data editor. 3. A custom program written specifically to handle the particular questionnaire. You can do 1 and 2 in R, but you can't do them as well as programs dedicated to those tasks, and you can't do 3 at all well. It depends a lot on the specific conventions of the platform you're working on. R is aimed at writing cross-platform programs, and isn't particularly good at writing GUI programs, which is what you want here. I would personally use Delphi for this, but there are lots of alternatives.
I usually use Epidata, available from http://www.epidata.dk/, which is free and enables you to write your own questionnaire file and then input your data. It also enables you to implement checks during data entry which in my view is an important feature. It enables export in a number of formats but its native output format, the .rec file, can be read into R using the foreign package.
Duncan Murdoch
Michael Dewey http://www.aghmed.fsnet.co.uk