It should make no difference to an objective examiner whether they are writing up a rejection or an allowance, nor whether the application is ultimately allowed or abandoned.
The truth of the matter, though, for me, is that it feels better to reject than to allow, simply because it's more conclusive.
(This has nothing to do with production requirements and whether or not I'm getting a count for what I'm doing.)
Don't get me wrong, though. I'm not talking about whipping up some bogus rejection based on crappy prior art that has nothing to do with the claimed invention, or making outlandish combinations of non-analogous art. I'm talking about solid rejections that clearly anticipate or obviate the claims. There is certainly nothing satisfying in denying an applicant a patent to which they are entitled, nor drawing out prosecution longer than is justified.
An examiner generally spends the bulk of their examination time searching for relevant prior art. This can take a whole lot of time, especially for the initial search, depending upon the invention being claimed and the examiner's familiarity with the claimed subject matter.
The thing is, once I find good prior art to apply, I know I'm done. There is a definite conclusion to the process (for the current set of claims), and I know that the invention, as currently claimed, is not deserving of a patent (subject, of course, to persuasive arguments, which does sometimes happen).
Making the decision to allow an application, though, is somewhat less conclusive. Sure, once I've done a thorough search and satisfied myself that I've found the best prior art available, I'm not reluctant to allow an application (OPQA be damned!)
Still, who knows what unavailable prior art might be out there? Some hardcopy of a journal that is sitting in some school's technical library in Boise? Perhaps an advertisement in a newspaper on microfiche in Berlin?
Actually, to tell you the truth, those possibilities don't even bother me that much when making an allowance. It's really the possibility that if I tweaked my search a bit, I might end up finding the mother lode of invalidating prior art out there somewhere on the Internet.
I've had instances in the past where the applicant had written their specification in such a way that it excluded some term that very well characterized their invention and was commonly used in the art. When that has happened, it has often taken a while and much searching and reading to find that term. But once found, the searching process suddenly got a whole lot more fruitful.
Obviously, you have to stop searching sometime. (The count system makes sure of that.) When you do, you just have to trust yourself that you've found the best art that is out there and allow the thing.
Even so, you never really know for sure. (There's a reason that 'Reasons for Allowance' gets phrased '...distinguishes over the prior art of record...', right?)
[One of the first questions I asked my SPE when I was new at the Office was "How do you know when to stop searching?". I never got a good answer. Now new examiners ask the same question of me, and I still don't have a good answer to offer.]
Rejections, when done correctly with good art (yes, let's emphasize that), are simply more definitive, more conclusive.
And that's why for me, they are more satisfying (particularly when it was a tough search to perform). To be sure, awarding a deserved patent is satisfying as well, particularly when you read all of the horror stories on the patent blogs about 'prosecution-from-hell' applications.
Still, knowing, without a doubt, that you've reached the correct conclusion simply is not possible when allowing an application; only when writing up a rejection can you really know that you've gotten it right.