The idea would be to see whether other Civil War researchers consider your correspondent's resources to be relatively good or authoritative, and whether your correspondent's method of research was fairly rigorous.
A database similar to what she's referring to is this:
http://www.civilwar.nps.gov/cwss/soldiers.cfm
It gives four Union Rodney Campbells, just as she said, one hit for a Confederate R Cambell (without the P), no hits for Rodney Campbell in the CS Army, one hit for Roderick Campbell, in the North Carolina Light Artillery. The only difference is it gives 202, rather than 175, hits for R. Campbell in the Confederate Army.
So it sounds like her database is similar to that, and in fact, they're both probably based on the standard place to go for Civil War soldiers, the records in the National Archives.
The records there certainly aren't 100% complete, but they're pretty good, and are the best compilation of records.
The problem is, of course, that Civil War records
are often findable,
if the person exists. If the person doesn't exist, well, how do you prove they aren't just missing from the records?
The US Census search through Ancestry.com lists two Rodney Campbells in the 1860 census, one in Pennsylvania and the other in Wisconsin, and the 1870 census is the same plus a child born in 1864. Ancestry.com's database of Civil War soldiers only lists three Rodney Campbells, all Union.
A search of the Official Records turns up no hits for that name. It's another big database, published in multiple volumes in the late 19th century, and contains military correspondence, which sometimes includes even privates cited for bravery or in some other unusual circumstance, though generally the names found would be officers.
In other words, if the option is open that the person might not have existed, that would be my conclusion.
If this were a normal case where I knew the person definitely did exist (someone was asking about their great great grandfather, for example), I'd guess that he either went by another name, by his initials, by another spelling, or there was some similar reason he wasn't showing up in the records. And that's what believers would always fall back on, of course. But honestly, if the ghost announced his name as Rodney Campbell, logic would indicate that that's also how he'd state it to the census taker, the recruiter, etc.
In other words, she picked a common-sounding name, a big battle, a common injury, and hoped for the best, but unfortunately luck wasn't with her.