I'm going to mirror what Faceng said.
I fought with those damn dewpoint errors until I was ready to tear my hair out. My Station (F1398) is properly sited, and it still borders on the edge of being a red x, even though my dewpoint numbers are accurate and a station 2 and 3 blocks from me mirror nearly the same.
What I found in my case, is that there is a long running station about a mile from me that CWOP seems to like to base it's "analysis" graph off of, and which follows PERFECTLY, but his station is on a roof eve, which isn't the best place. So the data is skewed...It also doesn't take into account there can be major differences in as little as a half mile, hills and valleys, etc. etc. etc.
So, don't worry about it. How he is performing the analysis and determines what is good and bad is unknown since it's not documented.
As long as your MADIS rating is above 90%, your good. Even near perfect stations I've seen get their data flagged occasionally here and there by MADIS. Forget the Gladstone analysis, I don't think he's done work on it in a long time either.
I think it can also be summed up as it's a helpful tool, but it's not "law". I'm assuming in all this your station is sited correctly and not next to any heat sources that can throw the readings off.