Preventing Learning Effects in Repeated experiments

[Ahmet Ilker Soydan ilkersoydan h]

Posted by ""Ahmet Ilker Soydan"" <ilkersoydan@hotmail.com>

Hello all,

I would like to know how to prevent the learning effect in repeated
experiments and see ""who does how"".

Let's say we are interested in 2 factors (A,B) on X (in a beer game setting)
and we design a series of experiment during which the factors A,B are
incremented in a uni-variate fashion, so that the effect would/may be due to
that factor (!). However, we know that the performance of the subjects may
be distorted due to learning which is not in the scope of the experiment.

I would appreciate if I could hear what are the common experimental
techniques to avoid learning effect (As you may have guessed, these
experiments are carried out with the same subject)

Thanks...

---------------------------------------------------------------------------
Ilker SOYDAN
Politecnico di Milano, Phd Candidate
Management, Economics and Industrial Engineering
Via Giuseppe Colombo 40, 20133, Milano
--------------------------------------------------------------------------
Posted by ""Ahmet Ilker Soydan"" <ilkersoydan@hotmail.com>
posting date Wed, 13 Jul 2005 14:22:46 +0200

[Weaver Elise A eweaver WPI.EDU]

Posted by ""Weaver, Elise A"" <eweaver@WPI.EDU>
Hello,

There are different methodological approaches to sequence effects:

1. Counterbalancing - this is the process of ensuring that different orders
of your effects are experienced by different experimental groups.

One form, the latin square, is described here:
http://www.uni-ulm.de/~cschmid/v1999w/kapit05a/k5a.htm

2. Between-subjects designs - this is the process that makes sure that
nobody gets any sequence effects.

3. Hybrid design - Some groups get the full sequence and others get part of
the sequence of measures. One example of this kind of design is the Solomon
Four Group Design, described here:
http://www.socialresearchmethods.net/kb/exphybrd.htm

All the best,

Elise
Posted by ""Weaver, Elise A"" <eweaver@WPI.EDU>
posting date Thu, 14 Jul 2005 12:35:05 -0400