quite a while ago, I ran into a problem with WU's use of the MADIS stations. It turns out that since there were a lot of them in my area, and they didn't report very often, my station (rapid fire) would report very different (but correct) temperatures during the early morning periods when the temperature rises rapidly. At the time WU had a quality control algorithm in place that compared stations measurements to an average of the surrounding stations. Well, you guessed it, in the situation I just described, it was MY station that got tossed out, even though it was more accurate and timely. Fortunately when I contacted WU about this, they quickly verified what was going on , and kindly set my station to "good" by default. They were nice enough to send me the actual output of their algorithm to verify what was happening, which I also appreciated. I'm not sure that they are applying any of the QC algorithms now...it actually doesn't seem so as I look around the area, bu tI don't know.
Anyway, it was not a huge problem, and it got resolved quickly, but since you brought up the MADIS site topic it reminded me of my little "encounter" with them.