6

I’ve been doing a lot of thinking on the Halo Effect lately. It seems that this effect can really make a difference in the outcome of my testing if I am not cognizant of its potential swaying power. How can I minimize its effect during usability testing on a web application?

  • 2
    Good question. To make it appear less self-promotional I took the liberty of changing the hyperlink so it leads to the relevant article on Wikipedia rather than to your own blog. It's also probably a better reference. – Vitaly Mijiritsky Jun 24 '14 at 7:12
  • Interesting. If a person puts in effort on their own blog. Why then would it be considered self promotion if they are merely referencing the work they put in? That's just Bizarre to say that it appears self-promotional. Plus, since Wikipedia is just a blog of sorts, what makes it the best new reference point. – eeklipzz Jul 1 '14 at 16:20
  • It's self-promotion because it's promoting your own blog. It has nothing to do with the amount of effort you put in it. Not to mention the SEO value of having SE point to it. Nobody said that Wikipedia is the best reference point, but it is a universally accepted, peer-reviewed, objective source. In this particular case it's also a more thorough article than the blog post, and a reference that you yourself cite in the post. – Vitaly Mijiritsky Jul 1 '14 at 16:40
7

Only wear dotted shirts when testing ;-) .

As described (in Wikipedia), the Halo effect affects "the observer", but in a testing situation, there's an observer (or moderator) and a tester. The observer is probably you (if you ask this question here), but I assume you are worrying that the tester is looking at some "stripes" and will report "wrong" results.

My next question is: are you worrying about whether the testing situation (invitation, expectations, payment, facilities, moderator, browser etc.) will influence the outcome? Or are you worrying that aspects of the tested software will influence it, which you don't want to test (e.g. testing interaction while your visuals have not been finalized)?

To address the first, there's a standard procedure which reduces the effect. Pay adequate compensation, dress similar to your participants, clearly explain what you expect (i.e., "We are not testing you, but we you are testing this software" - don't say "our software"), etc. I have no reference, but you can certainly find a book or website detailing professional testing procedures.

To address the second, you must be very clear what you questions are. Before testing, list research questions. Good ones might be "Do users find the Login button?" or "Is the product description legible and relevant?", a bad one is "Do user like the site?". Then design tasks which allow you to observe the answers. Good tasks are "Please log in to the site (given account and password)." or "Does product XYZ have feature ABC?".

There are also standard questionnaires in circulation, which allow you to rate the user's overall experience, after the test. Questionnaire building is an art, don't try to invent one yourself if you're not into psychology and statistics. A few examples are here.

1

The best way to eliminate the Halo Effect is experience. Undoubtedly you will make mistakes when starting out, as you study testing through books and pages and posts such as this one, and most importantly - continue to do testing, evaluating you own performance along the way - the more you will become aware of influencing the test. This awareness is the best cure for the Halo Effect.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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