[Sigia-l] card sorting: dealing with multiple placements

[Boniface Lau]

> From: sigia-l-admin at asis.org [mailto:sigia-l-admin at asis.org]On
> Behalf Of Sean Lawerence
>
[...]
> If you are looking to get a very precise result (a=0.5) through
> statistical analysis, then yes a behavioral or statistical
> background would be appropriate.

Statistics is not about precision. It is about understanding what the
data revealed regarding a phenomenon. In the process, different
methods were used to improve the confidence level on the revelation.


> If you are trying to complie users responses to an interaction to
> establish a baseline, perhaps someone with MLIS or a design
> background would suffice.

Compiling responses is a no-brainer. But all too often, people without
the proper training in statistics don't know how to properly interpret
the compiled data, let alone understanding the limitations of the statistics
method being used. Thus, people were fooled by the data that they
themselves compiled and wrong decisions were made.


Boniface