Our favorite subject, I know.
One of the biggest reasons many Lions fans support starting Stafford right away, even if it's still not clear that he is a better NFL Quarterback than Daunte Culpepper right now (yet) during the preseason, is that more starting experience will help Stafford become a better QB later on.
That is, let him start more games, make some rookie mistakes this year, and he'll be a better quarterback in 2010 and 2011, his second and third years.
Based on my research, I don't believe this is true.
History is relevant here, and while it's easy to get into an extremely anecdotal example/counterexample argument over stuff like this ("Peyton Manning started all 16 games." "Oh yeah, well so did David Carr, and Carson Palmer sat his entire rookie year" etc. etc.), it is a lot more useful to look at the entire trend for a large number of talented quarterbacks who were expected to start for the team. (By "talented" and "expected", I used first-rounder draft picks as the qualifier).
To do this, I went back to the most recent ten years' worth of first-round-drafted quarterbacks who have had at least three years in the league (that is, 1995 - 2006, as the 2007 and 2008 QB's haven't had much time to prove themselves yet). Note that I excluded 1996, as there were no first-round quarterbacks, and 1997, as the only first-round QB, Jim Druckenmiller, did not attempt a pass in his second or third year.
Here are all 28 recent first-round-drafted quarterbacks, with the number of Rookie STARTS on the X axis, and the AVERAGE of their second-year passer rating and third-year passer rating on the Y axis. Every gray point represents one quarterback. I'm too lazy to label all of them, but I'll include the list below.
Second- and Third-year passer rating, I believe, is the best measure to use. Presumably, by the time a quarterback gets into his fourth, fifth, and sixth years in the league, he's had so much experience that it really doesn't matter what happened way back when he was a rookie, so I'm only using years 2-3, the years Matthew Stafford is expected to really blossom into a good quarterback (please God, let it be so).
The red line is the linear average of the two variables (rookie starts and year 2-3 average passer rating). It represents the trend, given all these quarterbacks.
As you can see the red trend line is almost horizontal. This means there is a very slight negative correlation between rookie starts and year 2-3 passer rating (i.e. more starts as a rookie actually correlates to a worse performance later on). This correlation is very weak. Removing one quarterback would probably affect it significantly.
Starting more games in 2009 will not likely make Matthew Stafford a better quarterback in 2010 and 2011.
Starting Daunte Culpepper, if he turns out to be the best QB at the start of the season, will not likely hurt Stafford's development and his success next year and the year after.
Start the best QB on Sunday, even if it's the veteran. Stafford will be successful in the future. Go Lions.
(edit: the Y axis is mislabeled. It says "Year 1" and "Year 2", and it should read "Year 2" and "Year 3")