Hi,
I'm changing my database from SPSS to EpiData. IIt's a follow-up database, with almost 1050 subjects, and three different time points (baseline, 1fst and 3rd months). In total, I have 703 variables. I wonder if you can help me with which choice should I take to collect and merge this kind of data? Is it better to create only one database (including all time points), or should I have to create three databases - one for each time point? If I create 3 databases, how can I merge it? Ex: case n. 999 - merge baseline data with 1fst month, and with the 3rd month... Can you inform where I can find exemples? Thanks in advance.
Carlos
--- Hi,
I'm changing my database from SPSS to EpiData. IIt's a follow-up database, with almost 1050 subjects, and three different time points (baseline, 1fst and 3rd months). In total, I have 703 variables. I wonder if you can help me with which choice should I take to collect and merge this kind of data? Is it better to create only one database (including all time points), or should I have to create three databases - one for each time point? If I create 3 databases, how can I merge it? Ex: case n. 999 - merge baseline data with 1fst month, and with the 3rd month... Can you inform where I can find exemples? Thanks in advance.
Carlos
If your data are already in SPSS, it will be simplest to use the same structure as you already have (probably one database).
If you are going to be gathering data and using EpiData to input new data, then you have a choice. I won't offer a real answer, except to say that the choice of one database or RELATED databases will depend on how you want to analyze the data and whether there are differences in data points at the three visits.
If the three visits have exactly the same variables, you can use two related databases: 1. demographics with case number as KEY 2. visits, using case number as KEY to allow you to related the files (there is an example for this in classic Epidata - household, person, visit at http://epidata.dk/examples.php) In Analysis, read the visits file first and then merge demographics based on case number
If the three visits have different data, then you could have four related databases: 1. demographics as before 2. visit 1 with case number as KEY 3. visit 2 4. visit 3 This arrangement makes it possible to do the data entry independently, but still bring all of the data together for analysis. In classic EpiData, people did this to keep the number of variables down in each of the files. However, the data are structured exactly like a single large database. In Analysis, ready any file first and merge based on case number
You can make a single database and still make it easy to enter data by having a variable VISIT and depending on the value (1,2,3) Jump to the first variable for that visit. In the new EpiData (Manager, EntryClient) you can do this quite easily.
Jamie
On 2013-04-01, at 6:17 PM, epidata-list@lists.umanitoba.ca wrote:
I'm changing my database from SPSS to EpiData. IIt's a follow-up database, with almost 1050 subjects, and three different time points (baseline, 1fst and 3rd months). In total, I have 703 variables. I wonder if you can help me with which choice should I take to collect and merge this kind of data? Is it better to create only one database (including all time points), or should I have to create three databases - one for each time point? If I create 3 databases, how can I merge it? Ex: case n. 999 - merge baseline data with 1fst month, and with the 3rd month... Can you inform where I can find exemples? T
Jamie, thank you very much!
My SPSS databank is composed of 703 columns, where the first 200 are baseline, and the other the follow-up. Each line is a case. Every time I want to input data, I open the databank, search for the case number and for the situation (e.g. 3rd month). After this, I fill in with the data. My baseline variables are different from the follow-up. At follow-up, variables are the same, exept the variable identification: height1m and weight1m ---> height3m and weight3m.
In other words, your idea of making a single database, using a valiable VISIT and case number as KEY sounds great, but the VISIT variable I don't have in my original SPSS databank! How can I merge with this difference?
I suppose I'll continue with the case numbers: 1051, 1052, 1053, ... Is it correct?
I'm affraid if is (or not) possible to join the old data (from my 1050 subjects) with the new data from the EpiData, and how to do that? I must maintain the same variable names/description, otherwise I'll be completelly lost!
I was reviewing the example "Relate example - Household, Person, Visit", and it was very interesting for me, but I want (if possible) an exemple of how to merge SPSS databases with EpiData. Do you know one?
Once again, thank you very much!
Carlos.
2013/4/3 epidata-list@lists.umanitoba.ca
If your data are already in SPSS, it will be simplest to use the same structure as you already have (probably one database).
If you are going to be gathering data and using EpiData to input new data, then you have a choice. I won't offer a real answer, except to say that the choice of one database or RELATED databases will depend on how you want to analyze the data and whether there are differences in data points at the three visits.
If the three visits have exactly the same variables, you can use two related databases:
- demographics with case number as KEY
- visits, using case number as KEY to allow you to related the files
(there is an example for this in classic Epidata - household, person, visit at http://epidata.dk/examples.php) In Analysis, read the visits file first and then merge demographics based on case number
If the three visits have different data, then you could have four related databases:
- demographics as before
- visit 1 with case number as KEY
- visit 2
- visit 3
This arrangement makes it possible to do the data entry independently, but still bring all of the data together for analysis. In classic EpiData, people did this to keep the number of variables down in each of the files. However, the data are structured exactly like a single large database. In Analysis, ready any file first and merge based on case number
You can make a single database and still make it easy to enter data by having a variable VISIT and depending on the value (1,2,3) Jump to the first variable for that visit. In the new EpiData (Manager, EntryClient) you can do this quite easily.
Jamie
On 2013-04-01, at 6:17 PM, epidata-list@lists.umanitoba.ca wrote:
I'm changing my database from SPSS to EpiData. IIt's a follow-up database, with almost 1050 subjects, and three
different
time points (baseline, 1fst and 3rd months). In total, I have 703
variables.
I wonder if you can help me with which choice should I take to collect
and
merge this kind of data? Is it better to create only one database (including all time points), or should I have to create three databases - one for each time point? If I create 3 databases, how can I merge it? Ex: case n. 999 - merge baseline data with 1fst month, and with the 3rd month... Can you inform where I can find exemples? T
EpiData-list mailing list EpiData-list@lists.umanitoba.ca http://lists.umanitoba.ca/mailman/listinfo/epidata-list
The simplest way to get your data into EpiData is to export it from SPSS to stata (file type .dta), although this may or may not work with your version of SPSS
You can import stata files directly into either classic EpiData or the new EpiData Manager. If you import to Manager, you will likely have to check each variable to be sure the format is what you want (e.g. number of decimals). You can check this easily using the codebook function in classic epidata or one of the reports in Manager.
Then fix the fields you need to. Note that in Manager, you can select a group of similar fields to change the size of field (e.g. decimals or string length). If SPSS exports value labels, these should come into Manager OK.
a caveat - I have never tried the SPSS to Stata part; I have done Stata to EpiData.
The more complex, but guaranteed to get your data in exactly as you want is to export SPSS to tab-delimited. You then have to create your own epidata database and use import the data. This can get really messy as Epidata expects the data to conform EXACTLY to your specifications. I have had a lot of grief trying to get things to match.
You can also import that tab-delimited data into Excel (there are lots of web sites that explain how to do this). Then format each column in Excel to match what you want (e.g. number of decimal places). Make sure the last column has no missing values. I suggest adding a field called XEND with the number one in each cell. Make sure the first row has legal Epidata variable names. Select the entire table, copy the data. Then start up Epidata Analysis and give these commands:
read /cb save mydata
You have just created mydata.rec with all of your data. In fact, if you can select and copy all of your data in SPSS along with the variable names, there is no need for Excel. However, I think you may only be able to copy the data. You need the first line to be variable names for this to work. I have done this a lot (to get from Excel data to SPSS and it work well as long as the data are clean. Analysis expects all of your variables to be consistent and not have special characters (quotes, commas) embedded in text fields.
Good luck! Jamie
On 2013-04-03, at 9:42 AM, epidata-list@lists.umanitoba.ca wrote:
My SPSS databank is composed of 703 columns, where the first 200 are baseline, and the other the follow-up. Each line is a case. Every time I want to input data, I open the databank, search for the case number and for the situation (e.g. 3rd month). After this, I fill in with the data. My baseline variables are different from the follow-up. At follow-up, variables are the same, exept the variable identification: height1m and weight1m ---> height3m and weight3m.
Jamie, thank you VERY much! Well, I'll try first to export from SPSS. If I have problems, I'll try by STATA... This is my first time with EpiData, and everything is new. Thanks again, Best, Carlos.
2013/4/3 epidata-list@lists.umanitoba.ca
The simplest way to get your data into EpiData is to export it from SPSS to stata (file type .dta), although this may or may not work with your version of SPSS
You can import stata files directly into either classic EpiData or the new EpiData Manager. If you import to Manager, you will likely have to check each variable to be sure the format is what you want (e.g. number of decimals). You can check this easily using the codebook function in classic epidata or one of the reports in Manager.
Then fix the fields you need to. Note that in Manager, you can select a group of similar fields to change the size of field (e.g. decimals or string length). If SPSS exports value labels, these should come into Manager OK.
a caveat - I have never tried the SPSS to Stata part; I have done Stata to EpiData.
The more complex, but guaranteed to get your data in exactly as you want is to export SPSS to tab-delimited. You then have to create your own epidata database and use import the data. This can get really messy as Epidata expects the data to conform EXACTLY to your specifications. I have had a lot of grief trying to get things to match.
You can also import that tab-delimited data into Excel (there are lots of web sites that explain how to do this). Then format each column in Excel to match what you want (e.g. number of decimal places). Make sure the last column has no missing values. I suggest adding a field called XEND with the number one in each cell. Make sure the first row has legal Epidata variable names. Select the entire table, copy the data. Then start up Epidata Analysis and give these commands:
read /cb save mydata
You have just created mydata.rec with all of your data. In fact, if you can select and copy all of your data in SPSS along with the variable names, there is no need for Excel. However, I think you may only be able to copy the data. You need the first line to be variable names for this to work. I have done this a lot (to get from Excel data to SPSS and it work well as long as the data are clean. Analysis expects all of your variables to be consistent and not have special characters (quotes, commas) embedded in text fields.
Good luck! Jamie
On 2013-04-03, at 9:42 AM, epidata-list@lists.umanitoba.ca wrote:
My SPSS databank is composed of 703 columns, where the first 200 are baseline, and the other the follow-up. Each line is a case. Every time I want to input data, I open the databank, search for the case number and for the situation (e.g. 3rd month). After this, I fill in with the data. My baseline variables are different from the follow-up. At follow-up, variables are the same, exept the variable identification: height1m and weight1m ---> height3m and weight3m.
EpiData-list mailing list EpiData-list@lists.umanitoba.ca http://lists.umanitoba.ca/mailman/listinfo/epidata-list
Dear List, I have been using Epidata Entry 3.1 on my Windows Vista PC (from 2009). However, I have now switched to a MacBook Pro, on which I have installed Parallels Desktop 8 in order to run Windows and Epidata on the MacBook. However, I am unsure which Windows version to buy for this - should I use Vista or Windows 7? According to the info on Epidata Entry 3.1, I should not even be able to run Epidata Entry 3.1 on Vista. However, I have done this without problems even though the Danish introduction to Epidata (written 2008) states that the program is developed for Windows 95/98/NT/2000 (32 bit). So: Can I expect Epidata Entry 3.1 to run on Windows 7? Or should I buy Vista or something even older?
Sincerely
S. Andersen, m.d., ph.d.-student Denmark
I don't bother with Parallels - after all, you bought a Mac to escape from Windows. Classic epidata, including Analysis, runs fine using Wineskin on the Mac. I have Snow Leopard OS, but Wineskin is supposed to run well on Mountain Lion.
There is documentation here: http://epidata.info/dokuwiki/doku.php?id=community:mac_users_and_epidata
I think Epidata will run best under wineskin if you use a WS8 engine. I didn't put that in the wiki article yet. If you have problems with this, post it here.
If you use Parallels just for Epidata, install Windows XP.
I hope you have checked out the test versions of Epidata Manager/Entry Client on the Mac. The next release should be pretty solid.You will still need Wine or Parellels to run Analysis.
Jamie
On 2013-04-03, at 5:22 PM, epidata-list@lists.umanitoba.ca wrote:
I have been using Epidata Entry 3.1 on my Windows Vista PC (from 2009). However, I have now switched to a MacBook Pro, on which I have installed Parallels Desktop 8 in order to run Windows and Epidata on the MacBook. However, I am unsure which Windows version to buy for this - should I use Vista or Windows 7? According to the info on Epidata Entry 3.1, I should not even be able to run Epidata Entry 3.1 on Vista. However, I have done this without problems even though the Danish introduction to Epidata (written 2008) states that the program is developed for Windows 95/98/NT/2000 (32 bit). So: Can I expect Epidata Entry 3.1 to run on Windows 7? Or should I buy Vista or something even older?
Dear Jamie, thanks so much! I have read the wiki, and I wonder: It looks like Winebottler is a lot easier to handle than Wineskin - is there a downside to Winebottler compared to Wineskin? What is the difference?
Sincerely Stine
-----Oprindelig meddelelse----- Fra: epidata-list-bounces@lists.umanitoba.ca [mailto:epidata-list-bounces@lists.umanitoba.ca] På vegne af epidata-list@lists.umanitoba.ca Sendt: 4. april 2013 00:20 Til: epidata-list@lists.umanitoba.ca Emne: Re: [EpiData-list] Parallels, Windows and Epidata Entry 3.1
I don't bother with Parallels - after all, you bought a Mac to escape from Windows. Classic epidata, including Analysis, runs fine using Wineskin on the Mac. I have Snow Leopard OS, but Wineskin is supposed to run well on Mountain Lion.
There is documentation here: http://epidata.info/dokuwiki/doku.php?id=community:mac_users_and_epidata
I think Epidata will run best under wineskin if you use a WS8 engine. I didn't put that in the wiki article yet. If you have problems with this, post it here.
If you use Parallels just for Epidata, install Windows XP.
I hope you have checked out the test versions of Epidata Manager/Entry Client on the Mac. The next release should be pretty solid.You will still need Wine or Parellels to run Analysis.
Jamie
On 2013-04-03, at 5:22 PM, epidata-list@lists.umanitoba.ca wrote:
I have been using Epidata Entry 3.1 on my Windows Vista PC (from 2009). However, I have now switched to a MacBook Pro, on which I have installed Parallels Desktop 8 in order to run Windows and Epidata on the MacBook. However, I am unsure which Windows version to buy for this - should I use Vista or Windows 7? According to the info on Epidata Entry 3.1, I should not even be able to run Epidata Entry 3.1 on Vista. However, I have done this without problems even though the Danish introduction to Epidata (written 2008) states that the program is developed for Windows 95/98/NT/2000 (32 bit). So: Can I expect Epidata Entry 3.1 to run on Windows 7? Or should I buy Vista or something even older?
_______________________________________________ EpiData-list mailing list EpiData-list@lists.umanitoba.ca http://lists.umanitoba.ca/mailman/listinfo/epidata-list
The advantage of Wineskin is that you can put Epidata and Analysis into the same place, so that epidata files are all in the one virtrual C: drive. I found that running epidata was simpler with Wineskin, once it was set up. But with either one, you can keep your data files wherever you want.
Jamie
On 2013-04-04, at 2:32 AM, epidata-list@lists.umanitoba.ca wrote:
Dear Jamie, thanks so much! I have read the wiki, and I wonder: It looks like Winebottler is a lot easier to handle than Wineskin - is there a downside to Winebottler compared to Wineskin? What is the difference?
Sincerely Stine
participants (1)
-
epidata-list＠lists.umanitoba.ca