The title of the article is misleading, they are actually using a practical application mathematical statistics and probability.

Here is your math lesson for today.
They show the same hotels with the same prices on both PC and Mac lists, the difference is the
order in which they are shown. In other words, a hotel that is seven down on a PC list is three down on the Mac list. This is
unless you tell it to order cheapest first, then the order will be the same. The reason is statistical data gathered from the hotel reservations studied, as well as by Orbitz reservations.
*Side note: This would not really be considered bias data. They didn't care what type of room was rented, (they didn't have a vested interest) the room was already sold. They were looking for a pattern (Where do PC users tend to stay? Where do Mac users tend to stay?), using raw data. This is different than Wrigley saying that chewing gum helps math scores. They had a vested interest in the outcome, (wanted to prove gum was good) therefore it is bias.
What they found (according to the article) was that Mac users were more likely to stay at a more expensive hotel and even when they booked a cheaper hotel they were more likely to stay in a more expensive room. Therefore, Orbitz changed the order of the hotels/rooms, so it was more likely that Mac users would find the type of hotel and room they (other Mac users) tend to prefer. It is kinda like Amazon, Netflix, or Ebay with their section of you may like this also, because other people who are interested in the same stuff as you are looking at do. Of course those are not always right, just a Mac users won't all pick the more expensive places, but statistically it is more likely they will.