[EpiData-list] Visual (graphic) versus text definitions of data structures - please discuss principle and test EpiData Manager

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Tue Dec 29 14:19:22 CST 2009

Olivier Hacquet, Paris, France asked the question:
> I was wondering if it would be possible to edit qes file as in the 
> previous epidata entry? I think this function is useful to quickly 
> design a new data structure.
The principle for designing data structures with the new EpiData Manager 
(part "Define Data") are:
a: visual design of new structures with the mouse and/or function keys.
b: import structure (and data) from existing files of supported type 
(rec, qes, txt, dta, dbf, recxml ....)
        if the file contains data the number of records will be shown in 
the statusbar.

After one of a or b fields and labels can be moved with the mouse or 
change to default position by pressing F11, as well as fields can be 
renamed etc. Current data file stru ture is shown with F12 or 
appropriate toolbar button.

Obviously  users are in need of introduction and guides explaining this, 
but until we these exist the following notes can be used:

a. Visual design using the mouse and/or Function keys.

a1 Using the mouse:
--1 The user points selects with the mouse a "button" for a given type 
of variable on the toolbar
--2  The user moves the mouse pointer to a place on the "empty dataform"
--3  Clicks and a small question form is shown, where the name of the 
field and a field label can be given
--4  The field is shown on the form after the user presses "OK".

a2 Using only function keys and specifying field labels, names and details:
--1 press the appropriate function key for a given field type
      F2 (integer) F3 (float) F4 (string) F5 (text label) F6 - F8 (dates 
--2 The field will be added below the previous field without the use of 
the mouse.

a3 Using only functions keys (no field labels added, only field names, 
default types)
--1 press SHIFT+the appropriate function key for a given field type
      SHIFT+F2 (integer) SHIFT+F3 (float) SHIFT+F4 (string)  etc.

b. import structure from existing files of supported type (rec, qes, 
txt, dta, dbf ...)
In principle the same process is done regardless of whether the imported 
b1: Contains data (such as rec files, stata files, csv/txt delimited 
files etc)
b2: An empty qes file only describing the structure.

Therefore if the user likes the "QES" principle, then just create a qes 
file as before with any editor, save the text file and import or open 
the file after creating a new dataform ("Define Data" or Ctrl+N)  via 
the button "open" (Ctrl+O) or  add structure . 

Please discuss on this list whether there is still a need for a text 
editor, when the software can work with the "visual design" as described 
above. - and in particular when the possibility of importing a qes file 
exists. It only means that the qes file must be created outside the 
manager, which will be optimised for the new approach with the visual 
design. I think we could consider adding (import from clipboard), just 
as in EpiData Analysis, which in essense is the same as point b1 or b2 

 From an empirical point of view we should make some tests when teaching:
Take a questionnaire which is only on paper:
a - create a qes file and import that using point b2 above.
b - create the fields from the questionnaire and add fields using the 
approach a1-a3.
And in each e.g. measure the time to do so or the number of errors 
needing corrections for a class of say 20 students. Individual users can 
do the same test and compare the time.

Best regards and happy new year to all

Jens Lauritsen
EpiData Association

More information about the EpiData-list mailing list