Coding Not app, blanks, Don't knows

[leenasm]

I'm new to EpiInfo and am creating a questionnaire to input existing data about student academic performance. I've got a fair amount of missing data, don't knows, and not applicables. I'm planning to do simple statistical analyses on different subgroups of students. a few questions:

- Could anyone give me the "best practices" in Epi Info for coding "Don't Know", "Not Applicable", and "Missing"?
- When I use a Skip pattern, what does EpiInfo automatically put in the Access database for the "skipped" data? Is the cell just blank?
- I downloaded the EpiInfo6 Manual from the CDC website, but it appears to be very out of date (1996, doesn't mention "Make View"). Does anyone know of a more recent manual?

Thank you in advance!

Leena

[Biohazard]

1) It depends on what you want with your analysis. If there isn't a context for "not applicable", for example, then it shouldn't be included. Or if there is a context, and you just don't want to see those answers because you feel they are irrelevant to your analysis, then they can subsequently be left out of the available choices in the drop-down list. It's entirely up to you.

What I can tell you is that if you have seemingly redundant choices in your View, you can actually "recode" those several options into a single option later on in the Analysis module. I'll give you an example: Let's say you have 5000 patient records of people exposed to H1N1. If you want to do an analysis of sick people by age, do you do a frequency of age stratified by whether or not they are ill? Probably not - you would want to break those 90+ ages down into age ranges so you had a much more meaningful output. Thus, you would instead do a frequency of age range stratified by whether or not they were ill. This would give you maybe 5-10 rows in the output table instead of 90 rows.

And the same thing can be done here. So let's say you have Don't Know, N/A and Missing as choices or options. In Analysis, you can actually recode these to all just N/A or "missing" so that your output doesn't include three answers that all sort of mean similar things and aren't relevant to you. This way, the output will be easier to read, while at the same time, your data set has those extra values in case you actually do need them for something later on. But again, if you identify that there is no need for them, then it's just going to take you extra time to conduct that recoding.

I hope that wasn't too complex or hard to understand.

2) Any cell left blank in Epi Info is assigned a blank value in Access. The exception is the checkbox, which is either on or off and never has a blank or missing value. Contrary to often-times popular belief, a Yes/No field can actually have three values: Yes, No, and blank (missing), so it isn't actually the same as a checkbox.

3) The Epi 6 manual is for the MS-DOS version of Epi Info and is quite out of date. The Epi Info 3 manual isn't available in downloadable format. The next best thing to use is the Epi Info helpfile, which is the de-facto Epi 3 user manual. You can access it by going to Help > Contents from the main Epi Info menu screen. Make sure to use the SEARCH feature too, it can be very useful. Also, I'd check the Epi Info tutorials located at http://www.cdc.gov/epiinfo. They cover a lot of really good information and can be downloaded in PDF or DOC format. I highly recommend them for new users!

One last bit of information: If you ever get stuck in an Epi Info dialog box and don't know what it does, look for the Help button. Clicking that takes you to a help topic relevant to that function or feature.

Let us know if there's anything else we can do.