Hi Exline,
I always wanted to have a number between $10 and $15 so that it's more like a casual price that won't make people cringe to fork out their money. In my survey (http://marrily.com/survey), $10 - $15 is also the range that people are willing to pay as well.
A few response said they want to a price in the $5 - $10 rang, however, this price is too low, and in my opinion, it would affect user's perception of the application. They would think "you get what you pay for". Having a low $7.99 or $6.99 will probably hurt the app more than helping with cheaper price/higher volume. Based on my own experience (I was a paid customer to 37Signals basecamp for the longest time), at the end of the day, the difference between $7.99 and $12.79 isn't that much. I was willing to pay 37Signals $12 as much as I would if they priced it at $5.00. However, it's a huge difference in terms of revenue (a 37.5% loss in potential revenue).
I read somewhere that prices that end in 9, .99, and .79 work best. The Appstore probably an amazing demonstration of this: 99c apps do sell, and so do $9.99 apps for the iPad. However, I want to go with a price that works and not as as common. So I pick .79. And I'm superstitious as well, 13 is a bad number, so I don't want $13.79 (furthermore, (1+3+7+9) mod 10 = 0, not sure if it's good or bad here). $14.79 is a bit high (I probably have to do A/B split tests once the site got more tractions to test this price point though), while $11.79 is a bit on the low side. $12.79 is the perfect number.
However, all these come from a guy who has never sold anything in his life. So don't count on my words. I'm still in the experimenting phase, but I'd love to share my results afterward.
Cheers!
Edit: For 13.79, the sum of all the numbers mod 10 is 0, in baccarat, that's the worst hand you can possibly have. So definitely not $13.79 for the price.