Oops! Gliese 667 slipped through the cracks and wasn’t included in any of the original stellar datasets on my Stellar Mapping page! This is slightly embarrassing since it’s a bit famous for having planets around it! It was within 22.8 ly from Sol, but for some reason wasn’t on the RECONS list – and because it was so close it wasn’t included in the Hipparcos dataset either.
I’ve now added Gliese 667 to the RECONS CSV files, so if you’ve already downloaded the RECONS data, you’ll need to download the new version so you can include Gliese 667! (I only found it because I was checking the stars on the American Arm for 2300AD!). There shouldn’t be any other missing stars there – I checked the border between RECONS and HIP and couldn’t find any other HIP stars within 22.8 ly that weren’t already on the RECONS list.
Have you tried the formulas given at the bottom of the DENSE website for the weighted means and errors? I tried them on Lalande 21185, but couldn’t get the figures to match the ones given on the RECONS website.
I get the impression that those have already been calculated where necessary in the papers, so recalculating them would probably give you wrong answers.
I wanted to do a weighted mean on the parallax for Wolf 359. There’s one from 2009 listed on SIMBAD. RECONS is using the one from the 1995 GCTP. I suppose that just averaging the two would probably work. The one from 2009 put it further away. When I tried the weighted mean formula, it was even further. I also noticed the DENSE site had older parallaxes listed for some of the white dwarfs. The RECONS site is more likely to be updated, but I’d still check everything against SIMBAD and ARICNS.