[Sigia-l] obfuscating structure in card sorts?

[Eric Scheid]

On 18/7/04 3:52 PM, "Matthew deStwolinski" <matthew at destwo.net> wrote:

>> It doesn't take much imagination to rewrite these cards using synonyms
>> and different phrase structures...
> 
> Perhaps this is a sign that your fears are correct and that you've
> inflicted your bias on me, but it seems like you're trying to take sets
> of things that are very similar and manufacture differences that aren't
> there.

They are similar in different ways, and there may be similarities which are
not directly obvious. By phrasing each content label with the kind at the
front and the subject at the end puts the emphasis on the 'kind' of the
content, while some audiences might prefer that all the "method A" stuff was
on one page, the "method B" stuff on another. As for not directly obvious
groupings, that would require more familiarity with the context. It could be
that some things of A, B, and C would be grouped together, and the rest of
A, B, C would be grouped elsewhere ... because for those three things the
audience is interested in the benefits and the features and the pricing, and
not at all interested in the ROI, the how-to, and so on. Meanwhile, methods
D, E, and F might be split off to an "advanced stream" group, and for that
audience they may well be far more interested in comparing the ROIs of the
group.

Something else ... the actual content we're looking at isn't anywhere nearly
organised. There are some so called "case studies" which are closer to being
anecdotal narratives, some specific nodes don't yet exist, some exist in
three different versions, and so on. All quite a mess.

> I agree that you should try to keep your own mental structure of the
> information from biasing the subjects.  But by varying the descriptions
> as you've suggested, you run the risk of making this less about
> organizing items than about being able to correlate various
> terms/phrases together.  Are you more interested in finding out if users
> would lump all the cost/benefits analyses together separately from a
> lump of case studies?  Or are you more interested in finding out whether
> subjects know that cost-benefit analysis and ROI are related?

Related, but not necessarily the same thing, and that's the crux of the
matter.

e.