Part of the problem is being in a location where you can collect accurate data. The setup is very particular. You basically need an open field to do it properly. I would love to do it, but all of my landed is wooded and is a poor location for weather data collection.
There are actually some arguments about the accuracy of the automated data collected from the airports. Mainly stemming from the fact that airports are usually nowhere near the city in which they're collecting data for. There are also some dubious instrument configurations. For example, the setup at Reagan National, which is the official data for Washington, DC, was changed a couple years ago to where there are now rocks under the temperature station instead of grass. This is lead to some analysis that the official readings may be higher than the actual air temperature.
I'm not a volunteer observer, but I am the lead developer on the web app they now use to enter the observation data. The app is a result of the 'paperless' initiative the NWS (and NCDC) has. The person on the other end of his phone call is probably using this web app I work on.
