This is a common situation and with the limited information available it is really difficult to reach a scientifically valid conclusion about the accuracy of your RH data.
1) Since the other measurements are not right next door, you have the micro-climate question.
2) How accurate are the other stations? If your unit is +-7% error and the other units are (just for example) +-5% error then everyone could be "in spec" even if the readings differ by 12%.
Then the next obvious question is -- what are the odds that all these units are right at the edge of being out of spec? And at opposite ends of their spec limits? Then statistics rears its ugly head and you are faced with annoying questions like what are the mean and statistical distributions of errors in your sensor and the comparison units. And are there enouigh neighboring units for statistical comparisons to be valid? Yeck...probably not where you wanted to go with this.
It would be much simpler if there was a local source that had a known high level of accuracy -- like a local airport or NWS station of record. Then you have only the micro-climate question which might be resolved through comparison over long periods of time.
Or if you know anyone who has a more accurate RH gauge you could borrow and set outside for a while at your location.
If you think the height might be a issue, how difficult would it be to lower the sensor for a few weeks to check that? I don't know the answer to this myself -- I would think that unless the ground was really wet, 25 feet of elevation should not matter...but I don't really know.
I think you are right to be concerned about the differences between you and other stations; I've tried to show that figuring out why with any certainty is not so easy.
It might also help to consider how much error there is in the dew point calculations. You can find on-line calculators to that convert Temp/RH to dew point to play with, but here are a couple of examples.
T=60F, RH=40%, DP=36F. If RH error = +10%, then DP error = +5F
T=80F, RH=60%, DP=65F. If RH error = -10%, then DP error = -5F
So, even large RH errors translate to errors in dew point that are perhaps not as large as might be expected.
Lastly, let's say you figure out that you have a large error in your RH data. If it is outside the limits claimed by OS you can try to get a replacment sensor. If not then you may be tempted to apply an offset to your readings -- like say subtracting 7% from all the readings for example. This might or might not work as the actual offset may or may not be constant. It could vary over both temperature and humidity conditions. I'm not aware of anyone who's done this type of comparison over a wide range of temp/RH values to see how that would behave. It is tempting however and would obviously improve accuracy at least over some range of conditions. I've done this myself and while it makes me a bit nervous I justify it with the knowledge that at least the results are more accurate for some conditions.
