[LON-CAPA-users] automatically decrease point value with number of attempts

[Peter Riegler]

Hi Doug,

in my eyes the best way to get the desired functonality is to use 
spreadsheets. Years ago I had been using decreased point values, see
res/fhwf/riegler/Misc/Spreadsheets/
for the triple of spreadsheets I have used for this. If you want to try 
them it is a good idea to consult the Spreadsheet Help for the 
spreadsheet hierarchy of these three sheets.
Working with spreadsheets looks weird an complicated at first glance. 
But as always in loncapa, it is quite powerful and rational once you got 
the idea.
Frankly, I believe using spreadsheets is the only reasonable way to get 
the desired functionality. Problems should not be mixed up with grading 
rules. Mixing these up results in problems which fit a single instructor 
but are unlikely to be usable by others which defeats the whole idea of 
loncapa.

Peter

On 03.02.2015 19:54, Mills, Douglas G wrote:
> Hi All,
>
> I'm guessing there's probably a better way of doing this and someone on
> this list has probably done it.  Some of the instructors I work with
> would like to give, say, 5 attempts on a homework or quiz question but
> with decreased point value for each subsequent attempt, along the lines
> of 1point if you get it right the first try,  .8 points if correct on
> the second try,  .6 points on the third try, etc.
>
> Early one I tried the weight parameter with a variable for the value of
> it that changed with the number of tries. That worked beautifully… in
> the problem itself, but in the context of the assessment progress chart
> and other places in Lon-Capa, the variable was not evaluated so the
> weight turned out to be 0.
>
> Using "assigned_score" works well for this, but _as I understand it_
> requires use of the custom problem type. I've done a lot of these and
> appreciate their power and versatility, but they have their downsides,
> perhaps for me the most significant being the need to write code to
> evaluate significant figures and tolerance, which are done so well for
> me in numerical response problems.
>
> The approach used by some of the instructors I work with who do their
> own coding is to give the students access to the same quiz 3 times and
> then using the timestamps to determine which was the first attempt, the
> second attempt and the third attempt and plug those into excel to
> generate the best score with the second attempt worth .8 of the full
> points possible and the third attempt worth .6.  They have to look at
> timestamps because of course the clever and slightly unscrupulous
> students in the courses do these in reverse order to use the lower-value
> quizzes to practice for the full-value quiz.  I proposed using sequences
> to prevent access to the lower value quizzes until each of the
> subsequent quizzes has been attempted, but that has not caught on.
>
> So, is there a way, within one quiz, to allow multiple attempts at each
> question and automatically decrease the point value of that question for
> each attempt besides using "assign_score" (or is there a way to get
> tolerance and sig fig functionality into a customresponse without coding
> it yourself?)
>
> Or some other way to approach this that I'm missing?
>
> Thanks as always!
>
> Doug
>
> Douglas Mills
> Director of Instructional Technologies
> Department of Chemistry
> University of Illinois
> dmills at illinois.edu <mailto:dmills at illinois.edu>
> (217) 244-5739
>
>
>
> _______________________________________________
> LON-CAPA-users mailing list
> LON-CAPA-users at mail.lon-capa.org
> http://mail.lon-capa.org/mailman/listinfo/lon-capa-users
>

-- 
Peter Riegler
Ostfalia Hochschule für angewandte Wissenschaften
Hochschule Braunschweig/Wolfenbüttel
Salzdahlumer Str. 46/48
38302 Wolfenbüttel

Fakultät Informatik
Fon: ++49 5331 939 31540
http://ostfalia.de/pws/riegler

Zentrum für erfolgreiches Lehren und Lernen
Fon: ++49 5331 939 17200
http://www.ostfalia.de/zell