[Sugar-devel] Assessment in Karma

NoiseEHC NoiseEHC at freemail.hu
Wed Jul 29 06:01:18 EDT 2009


In June I was talking with the headmaster of a school which does 
education research. There are two things from that discussion that seems 
relevant for this question:

1. Karma could measure the time it takes to finish a task in a compound 
exercise. If it is statistically relevant then it could give some hint 
to the teacher which parts of the curriculum are hard for the given 
child. If I will be able to persuade the headmaster then there will be 
research (hard data) which hopefully will find out whether this 
measurement is meaningful or pointless.

2. There is a VERY HARD to develop exercise model where the child can 
follow multiple paths to the solution. In this case the possible choices 
the child can follow makes a graph with most of the leaves ending with 
"does not compute". If this exercise model proves to be useful (which 
will be determined in the mentioned research) then Karma should store 
the way the child solved the exercise and not only the end result. So 
what I wanted to say is that it is possible that the wrong steps can 
give more informations to the teacher about the knowledge of the child 
than the fact whether the child was able to solve the exercise or not.
Because the 2. point can be a little bit hard to understand, here is an 
example:
The problem is this: "If we used A amount of paint to fully paint a 
little picture with sides B and C, how much paint do we need to fully 
paint a room with sides D, E and height F?"
This is an exercise with 3 steps. The child has to calculate the area of 
the picture (P=BxC) and the walls' (W=(D+E)*F*2) and has to calculate A/P*W.
Now the graph looks something like this: from the beginning the child 
can calculate either area or can do something like B*E which "does not 
compute". If he has P and W then he can do either A/P or W/P then either 
A/P*W or W/P*A respectively. In every step the program should 
graphically show what the child just have calculated (what can be 
impossible if he uses D*E*F*B for example).

Now I do not insist that you develop this 2. point since it is a HUGE 
task (and needs some research to know whether it really help children 
understand complex relationships) but moving Karma into this direction 
(just a little bit) would help a lot if the headmaster would decide to 
do this research. In this case we will translate the resulting exercises 
to English, probably you will be interested in that.

ps:
I intend to use Karma for interactive curriculum development for reasons 
Bryan Berry talked about a lot.

ps2:
By research I mean "try out with kids while somebody from the University 
measures the outcome".


Christoph Derndorfer wrote:
> As previously mentioned by Bryan in his "Automated Assessment is the 
> Killer App" blog post 
> (http://karmaproject.wordpress.com/2009/07/06/assessment-is-the-killer-app/) 
> student assessment is an important component of Karma.
>
> While toying around with the lesson UI I realized that the assessment 
> area is still very much an empty space (which kinda reminds me of the 
> group view in Sugar;-) as we haven't really discussed how it should 
> look, what kind of data will be kept (which is of course heavily 
> dependent on the type of lesson), how the data will be stored, etc. To 
> keep the efforts required to do that in check I think that it might 
> make sense to start by offering a handful of templates - both on the 
> UI and storage backend side of things - to fit the most common 
> use-cases in terms of different lesson types (e.g.inserted words into 
> sentences, doing calculations, etc.)
>
> Now I was wondering whether anyone here had specific suggestions on 
> how to address this or pointers to how other e-learning solutions 
> (regardless of whether stand-alone applications or Web based) solve 
> this interesting challenge.
>
> Thanks in advance,
> Christoph
>
> -- 
> Christoph Derndorfer
> co-editor, olpcnews
> url: www.olpcnews.com <http://www.olpcnews.com>
> e-mail: christoph at olpcnews.com <mailto:christoph at olpcnews.com>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Sugar-devel mailing list
> Sugar-devel at lists.sugarlabs.org
> http://lists.sugarlabs.org/listinfo/sugar-devel
>   

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.sugarlabs.org/archive/sugar-devel/attachments/20090729/7e5b8f71/attachment.htm 


More information about the Sugar-devel mailing list