Mental models are extremely important in understanding how people approach a
product, what their expectations are, and in predicting how they are likely
to behave. For me, they are a key part of the design process. There is far
more to say about mental models then we can address here.
Mental models are a loosely defined term. They can encompass concepts,
feelings, intentions, beliefs, a person's perception of how things will
flow. I really understood the value of mental models when I encountered the
idea of a script. Let's say you decide to go to a restaurant. You will
(depending upon the type of restaurant) be greeted at the door, be placed at
a table, be shown a menu of choices, order food, be served food, eat the
food, pay and leave. When we enter a restaurant, we already have this mental
model and it organizes our behavior.
By understanding the mental model, we can build products that fit it, or we
can alert the user to the fact that his or her mental model may not be
accurate. You could say to customers on entering -- "this restaurant is a
fixed price no matter what you eat or how much" and that would cause their
mental models to shift.
But I believe it is better to think of both behavioral and cognitive
scenarios as being equally important; not in competition with the other.
Both need to be understood because behavior is the only thing we can
observe. Behavior allows us to validate the mental models we hypothesize
because if people do not behave as our models predict, then the models need
to be corrected.
I don't think the choice is between mental models and behavioral scenarios
but in the alignment and synthesis of both.