Thanks for the N/A explaination.
The Code i use is the latest (1.3 RC2) and the code in your CVS
repository is generating the 0 values for the ranking. Version 1.5 of
the survey_render.inc file.
Thanks again for the nice tool ;-)
CU
Olaf :-)
"James E. Flemer" wrote:
> Olaf, I am CC'ing the phpesp-general mailing list on this
> discussion, I would like for you to join the list and
> continue the discussion there.
> [ http://lists.sourceforge.net/lists/listinfo/phpesp-general ]
>
> On Wed, 9 Jan 2002, olaf wrote:
> > It looks like the 0 for the value is not correct. when
> > the result is calculated the avareg is made over all
> > values from that choice. But a 0 is not used
> > tocalculated (only ratings greater than 0 are used by
> > the select and average of 0 is also 0...) . When i
> > check only the 1. rank i never get a result.
>
> There was a bug in that. Try using the latest code from the
> CVS tree (http://sourceforge.net/cvs/?group_id=8956).
>
> > Another question :-) When i ue precision 1 for that
> > ranking, i get another row with N/A but in the result
> > these lines are not used (value is -1 in the
> > response_rank table) So is it planed to be in the
> > result or is it wrong that the N/A field is possible?
>
> Yeah, the precision option enables or disables the use of
> the "N/A" field (not applicable). If a respondant checks
> "N/A" then they are not counted in the "TOTAL" for that
> option. For example, the first respondant picks "3" for a
> question, and the second picks "2". The results now show
> "2.5" as the average (3+2)/2. Then the third picks "N/A",
> the result is *still* "2.5" because even though there are
> three responses, only the first two are considered because
> the third said that the question was not applicable (for
> whatever reason). However, as you have seen, the database
> keeps track of the number of "N/A" responses ... How would
> you suggest handling this data when showing the results.
>
> -James
|
