[LON-CAPA-users] surveys

Steven Pollock lon-capa-users@mail.lon-capa.org
Sun, 22 Aug 2004 21:11:07 -0600

I'm still working on getting a survey up. We've just upgraded to (today!) I wanted to wait for the upgrade before posting, but 
it hasn't seemed to fix these issues.

Guy's answer to my previous question about setting up a survey was very 
helpful, I feel close!(the goal being that I include both essay and 
multiple choice responses into the survey,
and then am looking to  give credit simply for their typing anything at 
all in)  But it's not quite all working.

> The survey style is what you want, this option curently works with
> <essayresponse> and <radiobuttonresponse>
> style questions.
> If the student answers the question it is marked internally as
> 'correct' and if any points are associated with it they are awarded.

This "any answer is correct" feature doesn't seem to be happening. I 
don't have any separate parts in my survey, but there are three 
"responses" (one essay and two radiobutton)
When I look at student grades, they don't seem to be getting any credit 
for having responded, it appears to still be ungraded.

> For <radiobuttonresponse>
> the 'correct' values are ignored, all foils are always shown and they
> are unrandomized (meaning the foils are shown in the order they exist
> in the problem)
> You can allow students to have multiple 'tries' in which case they can
> change their mind about the answer.
> Feel free to create a survey resource in CSTR (make a resource with an
> extension of .survey) and you can see how it acts.

The above all seems to accurately describe what happens!

> In version 1.2 (just coming out now) there are two ways of getting
> data about the answers out:

Does that mean I'm simply being premature trying this out in

> 1) in the STAT module you can download a summarized version of the
>    data  (I.e. 35 students picked option 1, etc.) (for essays all of
>    the submissions are shown/dowloaded)
>     There are no students names in this data, it is anonymous.

This works... although it seems a wee bit buggy? When I first set up 
the survey, I had my two radiobutton responses first, followed by the 
essay. When I got the report out, it was reporting out the
first responses as THOUGH they were essays (separate lines for each 
student, with some sort of numerical value for their button response 
appearing, appears to be 0 for the first foil, 1 for the second, etc.) 
The second and third responses were appearing as a "summary frequency 
table", which means the essay part was rather confusing.  (Just came 
out "100%")

But when I did nothing but change the order around, making the essay 
come first, followed by the two radiobuttons, it reported out fine: a 
list of all the essay answers showed up, followed by a proper summary 
(frequencies) for the next two button parts. (Very strange! )

> 2) in STAT you can choose to download the raw internal data for
>     problems as an Excel file, this does have student identifiers
>     associated with the data.

I cannot find this, (but it's precisely what I need.) I don't see this 
option in STAT survey reports. When I go to STAT student submission 
reports, only problems show up to be analyzed, the survey does not show 
Am I missing something obvious? Might I have somehow set up the survey 
incorrectly in the first place?

Thanks as always for feedback, much appreciated!

Steve Pollock

P.S. Our Lon-CAPA is "coming online" tomorrow for 2 classes, with over 
1000 students. Wish us luck!   :-)