Re: [EpiData-list] Re: Outbreak analysis diary - multiple data sets
Hi Jamie
Sorry I sent the message without finishing my section on multiple data sets.
Working with Multiple data sets in Epidata Analysis
What I was going to say was our first attempt at working with several copies of an outbreak dataset had been sucessful. This was a timing thing - we wanted to get the data in more quickly. We had already entered 150 records when we decided to use another data entry person, but that was Ok the process still worked.
Our steps were 1. Created a file and entered records (outbreak1) 2. Used Tools copy structure to create a new file (outbreak 2) 3. Entered records in both outbreak 1 and outbreak 2 4. Used Data in/out then Append/merge option B to create a new file outbreak_final, that was outbreak1 and outbreak 2 combined.
What was good about this was that it was easy, and I also really liked the option to copy the check file we had set up for the first file to use with the new file when you copied the structure. The only thing I didn't like so much was we ended up with a duplicate ID which we wouldn't get with an autoID (which we usually use for our ID field) but I used a number field rather than AutoID this time as I thought it would be a problem with the merge to have auto ID.
Using Multiple data sets in Epidata Analysis - READ and multiple data sets Could you explain a bit more about how to do this. When I try and READ more than one data set it says "no data will be lost close first". Then - what does the SAVEDATA do - create a new file like MERGE does? Which do you recommend to combine data - Merging in Epidata entry or in Analysis.
Using Epidata analysis in a field epidemilogy course I'm very interested in how you use Epidata Analysis in your field epidemiology course . Would I be able to find out a bit more about this from you - perhaps off the list to avoid taking up too much time on this topic? As I am involved in teaching courses using epidata/epi info to support field epi I am very interested in what you cover in your courses but realise others may not have the same interest
Many thanks
Christine
For multiple data sets (we do this every year in our Field Epidemiology Course), you can use Merge in EpiData Entry or just READ multiple data sets on one line in Analysis and then SAVEDATA.
Jamie
Christine wrote:
I really like the outbreak analysis feature but on my wish list would be
- for 'outbreak analysis' to be an option in one of the menus somewhere. I only knew about this feature through someone mentioning it on the list and I had to search through the docuentation to find it. Maybe this was what the 'attack rate plus RR' option was but I could not seem to select this - it stayed greyed out
Christine Roseveare Analyst Regional Public Health 570 9194 027 495 9671
1. Merge vs. Read and Append in Analysis
Apologies - the correct syntax to read multiple files in Analysis is:
read file1 append /file=file2
you can keep adding files.
SAVEDATA will save a new .rec file. This is especially useful if you have temporary variables that are used to manipulate the data. You can do this:
VAR DROP temp1 temp2 temp3 SAVEDATA finaldata.rec
Merging with Entry provides a bit of documentation of what was done and can be automated with EPIC (but thats another level of data management).
2. Merge and IDNUM. I normally avoid IDNUM as it is often useful to record ID numbers before you get to enter the data. However, you can use IDNUM and merge data successfully, if you assign each data entry person a unique range of IDs. In EpiData Entry, you can set an option for the first ID to use when entering the first record of a file. It is under File, Options and the Advanced tab. So one person could go with the default (1) and another could change this to 1001 before entering any data.
Another approach, where IDNUM was used and both sets of data start with 1 (although this is really not good data mangement practice, you might do this by including another code that indicates who entered the data). In this case, use Analysis to merge the data as above. Analysis doesn't care about unique ID numbers.
3. Field Epidemiology course and EpiData - we do a small field survey or observational study each year. We can send more info. Contact us via the e-mail address found at http://www.phac-aspc.gc.ca/cfep (Put my name in the subject line).
Jamie
participants (1)
-
epidata-list＠lists.umanitoba.ca