I've been designing radial menus (also known as pie menus, or contextual menus) for some time. Mostly for the video game industry. I've had a lot of success with them, and I find them helpful for presenting players with primary, secondary, and complimentary actions based on a single click.

But I've made a lot of assumptions with these menus, and I wonder if there is much research out there on the topic of hierarchy. For a menu with 4 choices, I've made the assumption that there is an order where the user (player) would look. Here is the order I've assumed:

1 Up

2 Down

3 Right

4 Left

Therefore, I've placed my actions appropriately:

  • Up - Primary
  • Down - Secondary
  • Right - Tertiary
  • Left - Least important

Do any of you know of any research to support/challenge this assumption? Is this just a heuristic call?

Thank you for the help!

  • 1
    I have no research, but I can tell you the order that I look in video games. Right -> Up -> Left -> Down, not sure why, but I'm sure thats how I read radial menus. Jan 14, 2013 at 19:35

1 Answer 1


No. Radial menus do not express hierarchy, and no menu item gets parsed first (the two are not the same, incidentally).

Research into radials (like, say, http://www.donhopkins.com/drupal/node/100 ) shows that seek time for any arbitrary element is generally consistent regardless of position. This isn't consistent with the premise that radial items are read in an order and that this order affects selection.

Another thing: don't assume that just because a user looks at certain elements before others, those elements enjoy primacy. There is a big, big difference between looking at something and parsing it. For instance, eye tracking studies show that in linear menus, users look at the first item, then the last, and then the rest. This implies that seek times on the last element should be very fast (second only to seek times on element 1) but they're actually usually the worst - because when the user initially scanned the first and last item, he did it to identify the bounds of the object, not identify the elements themselves.

  • I agree with some modifications: (1) Some research (krystiansamp.com/publications/samp_interact2011b.pdf) has found an effect of menu item position. (2) Population stereotypes exist for labeling a circle divided into quadrants. This is similar to ordering menu items in a 4-part pie menu. The expected ordering depends on the population. Perhaps there are common expectations among gamers. (3) However, the effect of menu item position disappears with practice.(billbuxton.com/MMUserLearn.html). Jan 15, 2013 at 16:49
  • @user1757436 - Hmm, that's some very interesting research. I hadn't seen those papers before. Did they appear in a particular journal? Jan 16, 2013 at 12:53
  • The first article was published in the Interact 2011 proceedings. The second was publishing in Human Factors in 1981. The third was published in the CHI 1994 proceedings. Jan 16, 2013 at 15:32

Your Answer

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

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