Motivation:

"Professor, I found the second edition on Amazon for 7 dollars.
 The third edition is being sold at the Student Store for over 80 dollars.
 Do you really think I should buy the third edition?"

Tis has been a recurrent question at the beginning of the semester.

The main course textbook for the class has a recently new edition. When deciding to buy the textbook for the class 
I tried to evaluate whether it was worth buying the newest edition. 

Whas it essential. The changes since last edition were hard to find at first glance, and moreover,
 if I did find changes they could have been easily published as 
additional accompanying materials (this was actually the case for
 this textbook). The improvements from the previous
 edition included more great examples, more cases
 that related to the textbook material.
 All those could be added to a book-accompanying web-site,
 for example (and indeed this was done by the authors of the textbook). 

From this and previous experience , I suspect that in most cases it does not justify a complete new edition.
It may very well be the case that publishing firms drive up the cost of books by using tactics such as bundling textbooks with CD-ROMs, extra problems or detailed solutions to exercises. These new editions may be published without making significant content changes to the body of the text.
 
More importantly, It is very much possible that, in some cases, the content changes do not justify the price differential from both old and newest editions (with respect to above numerical (true) example, even controlling for the fact that the second edition is probably sold (used) at Amazon and the third edition is sold new at a brick-store ). 

I did a very simple exercise and compared the prices at Amazon between successive editions for famous textbooks in two major fields of the social sciences. I divided the price of the new edition by the price of the previous edition. Each price should additionally be divided by the U.S. GDP/per capita of each year of release. All comparisons are between list prices of same type books (that is, hardcover with hardcover). 
What I found with this small sample, is that these price differences can be on average in the order of magnitude of 2.52 (although standard error is 1.5).
 (Source: calculations in twenty minutes of on-line surfing, with a very sample of 8 textbooks. The books with lower price differentials were core course textbooks for a particular major, the ones with higher price differentials were for elective textbooks. For one book, not included in sample for the reason to follow, previous editions were not available at all. this is the case for a lot of textbooks. Among the books in core courses, the highest edition was a 5th edition, while in elective was a sixteenth edition with CD-Rom). 

It would be extremely interesting to perform a more careful study on this issue.

If indeed the high cost of textbooks is contributing to the rising cost of college, it makes it more and more difficult for students to afford college. If the rise in cost of textbooks is not justifiable with changes in demand conditions (rise in income for example) and changes in supply conditions, then this is an important question.