I have my copy of Inside SQL 2000 at my client's office where I won't be until thursday, so I can't check exactly what it says there. However, like I said, if you compare these queries you will see that they use almost the same execution plan.

Usually a join is more efficient than the alternative.

I usually look at what is most efficient in a specific situation.

If you use alot of corrolated [sic] subqueries you may have to work on thinking more set based, and less row-by-row.

Why is a correlated subquery row-by-row based?