Logical?

The point is that without any other evidence you can't know which details are correct and which ones aren't. And sometimes evidence that appears to confirm a detail merely incorporates the same error, as in the case of someone who marries twice but gives the same ficticious name for their father.

There is so much scope for error in marriage entries after 1837. Either or both parties could give the wrong age, or claim to be 'of full age' when they're not. Even if the father's name is correct the occupation might be wrong: someone out of work could be described as a gentleman, an ag lab might be upgraded to farmer. There's also the matter of where the couple were living: research by Professor Probert has found that even if they gave the same address it didn't mean they were co-habiting. Middle names may not appear on the individual's birth certificate, or in the baptism register (or vice versa).

We're so used to these small discrepancies that we often don't notice them, but less experienced researchers might put great store by them.

The number of instances you found in a single parish might be exceptional, but this type of error is far from unusual. It's actually easier to diagnose these errors when there are a lot of them - when there are, or may be, isolated cases they're much more difficult to detect. Family reconstitution can help, but it's a time-consuming process.