It seems that Sarah's technical problem has been resolved. But it leads me to a more general remark about the handling of text responses.
If you are working with questionnaires with text responses it is smarter to code responses before entering the data. You might have a prior idea of which categories are interesting, or you may write down a list of responses for a sample of 100 questionnaires, organize the list, and decide the coding categories.
Next, write the numerical code on each questionnaire. Enter the numerical code, not the text, in the computer. This step is important because:
1) Few human brains are good at combining decisions (coding) with routine tasks (data entry); it slows down the process and you make errors.
2) If your data are important and you want to enter data twice to avoid errors, you will get a mess if responses are not coded before data entry.
I recall a dataset where the investigator for some reason wanted the record not only the smoking habits, but also the brand of cigarettes smoked; it was entered as text. It turned out that Danish women can spell Marlboro in 163 different ways (was my spelling correct?), and the recoding to something meaningful was quite a task.
Just some general thoughts; obviously Sarah can't use them for her current problem.
Svend
________________________________________________________
Svend Juul
Institut for Folkesundhed, Afdeling for Epidemiologi
(Institute of Public Health, Department of Epidemiology)
Vennelyst Boulevard 6
DK-8000 Aarhus C, Denmark
Phone, work: +45 8942 6090
Phone, home: +45 8693 7796
Fax: +45 8613 1580
E-mail: [email protected]
_________________________________________________________
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
