You are quite right. My only point is that Terrell, who was supposedly more moderate and electable than the conservative Woody Jenkins actually did significantly worse in the rest of the state, although she did slightly better in Orleans Parish. Some of this was due to Landrieu's so-called "sugar issue" in which socially conservative, normally Republican voters in Acadiana abandoned their moral beliefs to support a protectionist position.
The other nail in Terrell's coffin was John Cooksey's refusal to back her, which cost her dearly in the normally Republican 5th District. Republican turnout in this district was further depressed because of bitter feelings in the 5th District Congressional primary.
Many factors combined to doom Terrell. But they had one thing in common. They resulted in a much smaller plurality for Terrell among white, normally Republican voters, ie-Landrieu got a larger share of normally Republican voters in 2002 than she did in 1996. This would have had to be true because her plurality in Orleans Parish was 20,000 votes less in 2002 than it had been in 1996. If Terrell had gotten the same share of white, normally Republican voters in 2002 that Jenkins got in 1996, she would have won. Unfortunately, much as we might like to blame Orleans Parish for Landrieu's win, it was a failure of turnout on the part of the Republicans.
I think a lot of that is attributable to the election dynamics of presidential year versus non presidential year, even when the race being a runoff is taken into account. Presidential years typically raise the profile of elections in general and prompt voters to return a few weeks later in the event of a runoff (this also happened in '96 for the runoffs in some of those congressional districts that were altered in the wake of the Vera case).