College completion puzzle (southern states edition)

Does Alabama have 14 percent or 32 percent college completion among young adults?  The State of the South 2010 report offers both numbers, just four pages apart, each time reflecting both associate's and bachelor's degrees.   For Kentucky, the choice is 18 percent or 34 percent, and the numbers for all states are shown below:

The pipeline column shows data presented using the familiar method, pioneered by Tom Mortenson, that begins with "Out of 100 ninth graders" and works through who completes high school, goes directly to college, and finishes college within 150 percent of expected time.

The adults 25 to 34 column shows data taken from the Census Bureau, and the final column shows the difference that I think genuinely deserves to be puzzled over.

Since I've been puzzling over the mismatched numbers for several years, I'll share the best clues I've found.  I think the pipeline method misses key issues in its first and last steps:

• At the beginning, it uses the state's total reported ninth grade enrollment, including those repeating the grade. A student who is held back is counted as enrolled in two different years, but can only be counted as graduating from high school once.
• At the end, it uses each state's reported college completion rate, which is the percent of full-time students who graduate from the school where they first enrolled.   That means that students who transfer at any point in their undergraduate careers are not counted as graduating, no matter how quickly they actually complete their degrees. 

Without saying that either column offers results I like, I think the difference matters: two respected sources produce quite different evidence about educational outcomes for young adults across the South.