Hi All,
Jens was here at Epicentre to discuss some of the future plans for Epidata, and we had a brief overview of Epidata manager. As someone who writes qes/chk files using a text editor, I was disappointed to hear that it would not be possible to do this using Epidata manager (although the possibilty is there to import pre written qes files). In discussion with Jens, we talked about a solution to how a pre-written chk file can also be incorporated with Epidata manager.
My suggestion is to have a data dictionary, where you not only specify the variable name, variable label, value labels and the data format, but you also add attributes for missing values, range checks , legal values etc. The dictionary could be a CSV file, that can be easily created using excel.
Something along the lines of .....
Ideally however, what I would like is to have more flexibility, where you can assign the same attribute to the variables concerned in one statement, rather than listing one by one as a above, for example For all variables whose value 9 should be treated as missing, the attribute missing can given using for example the for loop as in stata,
for gender locat type case { missing=9 }
Also for those variables that share the same labels, you should be able to define one label and then assign to the variable concerned.
Once the dictionary is imported into Epidata manager, you can then use the graphics device to organise the data entry screen.
Food for thought, I hope? All the best Sarala
11 March 2010
Hello All, I follow the plans and see a lot of work coming for Jens and his team and despite that all I still cannot see the gain in it. When users are too lazy to learn a bit typing to make the necessary RECfiles and CHECKfiles ,these users can never be good EPIDEMIOLOGISTS. MOUSERS should stick to BROWSERS. Louk
Hi All,
Jens was here at Epicentre to discuss some of the future plans for Epidata, and we had a brief overview of Epidata manager. As someone who writes qes/chk files using a text editor, I was disappointed to hear that it would not be possible to do this using Epidata manager (although the possibilty is there to import pre written qes files). In discussion with Jens, we talked about a solution to how a pre-written chk file can also be incorporated with Epidata manager.
My suggestion is to have a data dictionary, where you not only specify the variable name, variable label, value labels and the data format, but you also add attributes for missing values, range checks , legal values etc. The dictionary could be a CSV file, that can be easily created using excel.
Something along the lines of .....
Ideally however, what I would like is to have more flexibility, where you can assign the same attribute to the variables concerned in one statement, rather than listing one by one as a above, for example For all variables whose value 9 should be treated as missing, the attribute missing can given using for example the for loop as in stata,
for gender locat type case { missing=9 }
Also for those variables that share the same labels, you should be able to define one label and then assign to the variable concerned.
Once the dictionary is imported into Epidata manager, you can then use the graphics device to organise the data entry screen.
Food for thought, I hope? All the best Sarala
Certainly, many of us will want to continue to have a simple way to create and edit checks (and .qes). With the shift to a clearly defined XML format for EpiData, this should be possible using tools we already have - that is, we could just edit the .recxml file, using cut and paste to replicate checks for similar fields.
A great addon or standalone module will be a simple XML editor for recxml files, to be used at our own risk. I don't know if such things exist in other areas, but there must be.
Importing a table of attributes as proposed by Sarala would be a great addon that could come from the user community once the .recxml format is finalized. Similarly, a utility to merge additional xml attributes to a .recxml would be very simple and very useful.
"USE fieldname" could be included in any of these, and probably the suggested "for var1 var2 ..." approach.
Jamie Hockin Ottawa
On 11 March 2010 Sarala wrote:
In discussion with Jens, we talked about a solution to how a pre-written chk file can also be incorporated with Epidata manager.
My suggestion is to have a data dictionary, where you not only specify the variable name, variable label, value labels and the data format, but you also add attributes for missing values, range checks , legal values etc. The dictionary could be a CSV file, that can be easily created using excel.
Ideally however, what I would like is to have more flexibility, where you can assign the same attribute to the variables concerned in one statement, rather than listing one by one as a above, for example For all variables whose value 9 should be treated as missing, the attribute missing can given using for example the for loop as in stata,
for gender locat type case { missing=9 }
Also for those variables that share the same labels, you should be able to define one label and then assign to the variable concerned.
Once the dictionary is imported into Epidata manager, you can then use the graphics device to organise the data entry screen.
epidata-list@lists.umanitoba.ca
-
epidata-list＠lists.umanitoba.ca