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