If you have a bunch of mental models derived from card sort and user interview data from a single user group, is it okay to attempt to form a universal mental model by looking at patterns in the different models and forming decisions based on the overall consensus? Or, is this in violation of this citation from Steven Heim?

Mental models are:

Personal—They are specific to each individual and are not universal concepts that can be appliied generically.

The Resonant Interface, Steven Heim

Note: I never use mental models verbatim in future decision making of site maps and wireframes; intuition and usability testing plays a large part in getting it right.

  • 1
    Please ask a question that can be somewhat objectively answered with a "best" answer. The way you've worded the question at the moment invites numerous inputs from people based on their personal experiences and opinions, which isn't really what this site was designed for. If you want to discuss approaches to mental models, try our chat room!
    – Rahul
    Commented Jan 22, 2012 at 2:25
  • Can you clarify this " I ask participants to create their own mental models once they have grouped and labelled the deck."? I haven't come across using this concretely, more as a general concept.
    – Illotus
    Commented Jan 22, 2012 at 9:13
  • Thanks for the feedback @Rahul, this has been a difficult one to try and communicate. I have attempted to reframe this question to elicit less discussion and allow for a “best” answer. Further feedback welcome.
    – DigiKev
    Commented Jan 22, 2012 at 15:31
  • Great improvement. I removed the "previous version" because it detracts from the rephrased question you want to ask.
    – Rahul
    Commented Jan 22, 2012 at 17:39

3 Answers 3


Mental models are naturally very nearly unique for each person. However large parts of those mental models are shared by people who come from the same (culture) group. Forming a consensus based on bunch of representative mental models is basically what (evidence-based) UX/usability research is all about. Most good for most people among the users of the product/service (barring any need to concentrate on servicing certain user group).


People do share 'high level' mental models (or schemas)

My mental model of say 'house' and your mental model of 'house' have to be generally similar - otherwise nobody in the world would be able to understand anyone else.

Here's a (fairly simplistic) link to a description of schemas:


Personally, I wouldn't expand the model outside the initial scope.

The thought and analysis that went into setting up the initial card-sort, and how the follow-up questions were phrased or grouped would have been done with specific, researched guidelines that could skew the results if applied across disciplines. I wouldn't even apply it to a broader group which the initial participants were a sub-set of. The variables that caused them to be a sub-set would work against a more universal model.

To be generic enough to apply universally, I'd think the results would be too broad to have meaning if extrapolated, and it's then educated guess work.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.