Take the 2-minute tour ×
User Experience Stack Exchange is a question and answer site for user experience researchers and experts. It's 100% free, no registration required.

I am looking for research into user acceptance of suggestions that they might receive from a decision support engine (spam filter, for example). Are users more inclined to take the suggestion if it aligns with their expectations? If the suggestion does not align with their expectations, how best to 'persuade' them to comply with it?

I have searched various scientific literature, but I only see bits and pieces here and there. Does anyone know of a comprehensive study or a review paper that covers these topics?

share|improve this question
    
Assimilation contrast theory, Havland et al. en.wikipedia.org/wiki/Social_judgment_theory has links to Research grade materials. If you are wanting to influence people, you might also read up on 5 sources of power (French, Raven) and subsequent derivatives. –  rlb Nov 12 '13 at 11:12
1  
In the human factors literature, you would search for 'trust, and 'automation'. This is a survey of the existing research as of 2004. This is a much shorter survey from 2010. This is 2013 view of this topic from AI researchers. –  user1757436 Nov 18 '13 at 17:17
add comment

1 Answer

It's kind of long, but it might be worth reading the whole thing. Specially the Experiment section, it's fun. :)

Background Info:

In a TED video by Shyam Sankar he suggests that:

You can start by designing the human into the process. Instead of thinking what a computer would do to solve the problem, design the solution around what the human would do as well. When you do this you will realize that you'll spend all of your time on the interface between man and machine. Specifically on designing around friction and interaction. In fact the friction is more important than the power of the man or the power of the machine in determining over all capability."

Analytic Capability Equation

Where :

  • a = Analytic Capability
  • h = human
  • c = computer
  • M(h*c) = Gestalt of (Human & Computer)
  • 1+fi = Friction

He concludes saying that:

Computers don't detect novel patterns and new behavior, but humans do. Humans using technology, testing hypothesis, searching for insight by asking machines to do things for them.

And to test an hypothesis you've got to have an expectation first and search for insight, this way, the mind is already ahead of the request made to the machine (or decision support system).


Experiment:

Why is Gestalt of Human and Computer (M(h*c)) embedded in the Analytic Capability?

Well, According to Gestalt Theorists:

the ability to perceive objects [or find results] on a screen is a result of prior knowledge and expectations + image on retina.

In your case an answer or report or solution from the decision support system.

Let's make an experiment. See the following image and try to identify things in it. Something. Anything:

Ambiguous Stimuli Image

When presented with ambiguous stimuli, our knowledge of the world helps us to make sense of it -- same with ambiguous info on computer screen. At this point you have no prescribed expectation of what to find. Hence, the image is likely to be almost senseless.

From my experience in the office, me and some coworkers were clueless to what was represented in the image, then we were given a little bit more time, and nothing.

This would be similar to a decision support system, giving us numbers without units, headers, titles, boundaries, etc.

Now, look at the image again, and search for a Dalmatian. Aha! There it is! Same image, different you (a you with the expectation of finding a Dalmatian).


Answer:

Are users more inclined to take the suggestion if it aligns with their expectations?

Yes, as expectations do bias result perceived quality.

If the suggestion does not align with their expectations, how best to 'persuade' them to comply with it?

Educating your users to align their expectations to the results your decision support system delivers.


Sources:

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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