As I mentioned on Saturday, I strongly encourage junior people to try and get on an NSF panel. A couple of people commented that they have tried to get on a panel and not been successful. Although this has not been my experience, I can see how this could easily be the case. As I already stated, POs need to balance a number of things when putting together a panel. If your research topic is smack in the wheelhouse for a particular panel, there may be many willing participants who fill your "panel niche" and it may be tough for you to get into the rotation.
What surprised me a little was Zen's comments:
As solid as this advice is, it bugs me. It makes it look like getting an NSF grant is determined by unwritten standards maintained by an established network of well-connected people, with fairly high barriers to entry.
The NSF shouldn't a be a country club....
....That there is a substantial number of replies indicating people want to serve, but haven't been invited, and that there's no clear understanding of why that is, does not help promote the appearance that panel selection is a transparent process.
This line of thinking doesn't make sense to me. How, exactly, would you make panel invites a transparent process? Considering that NSF heavily guards it's panel membership, I don't see how this goal is attainable. I'm not sure it should be.
The reality is that some people make better panelists than others, based on their experiences. Maybe the PO has seen your proposals and likes what they see. Maybe you tick the right boxes in terms of the knowledge that a panel needs to fill a hole. If it were just first come first serve then there would likely be large gaps in the field coverage of the panel. When you account for the need to include gender, ethnicity and employment stage/place balance, yes, there will be some people who have a hard time getting on a panel.
There's nothing Country Club or In Crowd about it. I did not know any of the POs on my first panel and my training was in another country, further reducing my connectivity. I contacted my PO after an unsuccessful application and asked to serve on the next panel while revising my proposal. For me, it worked. For others it may not be that simple, depending on their expertise and the needs of the panel.
I don't see how the selection process could or should be a "transparent" process, but maybe there are some arguments out there.