I'm re-reading Dan Ariely's Predictably Irrational right now. It's still amazing to me how completely manipulable we are. I just finished the first chapter, which deals with the phenomenon of how our decision-making processes are so predictably skewed by changes to the choice pool.
The example Dan brings up in the book is an experiment where photos are taken of students who are comparably attractive. Without a third option, giving the photos to students and asking them to choose which subject of the photo is more attractive returns a result relatively close to 50%. After a little Photoshop slight of hand, a third option is created by using the orignal photos and adding some imperfections to the faces. Now, presented with three choices (let's call them A, B, and not-so-B), the results skew toward B in a consistent manner. Likewise, if the choice pool is A, not-so-A, and B, the results skew toward A.
What's going on is that, despite our rational minds, we see the similar-but-better choice as desirable (even as compared to the non-similar option). When we have a basis of comparison ("B is like not-so-B, but better") it is easier for us to decide than if we are comparing unlike things ("A or B... they're both pretty good."). As a result, the "Be or not-so-B" is a much easier decision to make than "A or B." so our irrational minds make the leap to "If A and B are similar, but B is clearly better than not-so-B, then B must be the correct choice."
You can see this in play outside of the lab, as well. When pitching ideas to a group, it's fairly standard practice to offer three options: good, better, and best. Generally, the "best" option is the one you want them to choose. It doesn't always work out that way, but that's the goal. If you take Dan's lesson to heart, the way to guarantee that they select the option you want them to would be to make better similar to best, but a little less desirable. If you leave too much of a difference in quality between all three choices, the group would lose that ability to pair and compare two of the choices.
When you apply this to other real-world situations, you can see how it can be made to affect pricing structures, product selection, etc.. And if you're still dating, you may want to find someone who looks kind of like you, but it just a little less attractive. Trust us, you'll be happy you did.