Was it the same tester both times at the wellhead? Was the same methodology used both times? Is it possible the method the first time was performed poorly?

I've seen large discrepancies based on sampling differences. The Air Chek method keeps the sample running at a very low speed, without the chance of aeration, whereas the"bucket method" of running the waste into a bucket and dipping the vial into the bucket can yield wildly varying results.
There's the additional question of inline flitration. I've seen outdoor water run through a charcoal filter, instead of the indoor. Somebody installed them opposite of what they should have been.
I would think if the wellhead sample was actually taken from the well itself, not a well local spigot, if the sample was taken from the top foot or so of the well, the radon would likely have outgassed, due to exposure to the air within the well.
The right aquifer can vary wildly depending on the well level too. If the well is more full and that happens to be where the source of radon is, then you'll see a much higher reading. Likewise in that same well, if wash has been done, showers taken, cars washed out maybe some well meaning person bleached the well, then flushed it out to get the bleach out and the well hadn't recovered yet, and the radon was in that upper level of the well, you may see much lower radon levels.
I'm sure there's many other differences than can cause that variability, these are just a few that I've seen personally.

Yes to all.

"The right aquifer can vary wildly depending on the well level too. If the well is more full and that happens to be where the source of radon is, then you'll see a much higher reading."

I was wondering about that as a possibility too. If the water level inside the well was quite different between the two sampling periods, maybe the water level was below the source for the lower radon reading.

Then again, how deep into a well is a tester really able to go? I don't know anything about wellhead sampling, I'm guessing they don't go down very far.

That's why, even though the AirChek radon in water sample kit is designed to take the water from a spigot, we have adapters, which can be bought at any hardware store, to take the water from an indoor faucet. That's where our point of use is, that's where the sample should be taken. It seems to minimize most of the factors that can make the readings vary so wildly, with the exception of course of varying well levels.
This also eliminated the discussion I had with home owners early on,
"the outside spigot is not the water we drink from THE well, why are you taking the sample there? "
They were absolutely correct so we adapted. Even though they only had 1 well. The outside spigot, if the water in the house is filtered, is typically not filtered. It extends the life of the filters.

30K and 169K Samples were all taken from hose faucet on wellhead after running water for 10-15 minutes. Thanks

Simple explanation is sometimes the most likely, if sample had one tiny bubble of air inside it would radon would escape at lab when tech opened the bottle to take sample for examination. I always do duplicate or triplicate sampling and all labs are not equal in testing for radon in water. It is not worth my expenses for having to go back for free when I hit a weird sample result. That said you probably do many more tests than I do. As I tend to be more expensive because of my testing protocols.

Happy new year to everyone

Valid point. I do duplicate samples, and they were within acceptable RPD. Thanks,

This is 30 year old data is from 10 wells in Maine measured over 3 weeks showing a variability of about 1/2 to double over 3 weeks. Of course this is Maine.
Your results are a lot greater. Getting a precise water sample takes a precise method for sure. In my last research I used a hose connection at the indoor well tank faucet routed into a 5 gallon bucket. I set volume low on the hose and when the there was enough water in the bottom of the bucket I placed hose into the sample vial under water and then capped it under water. Then repeated with a second vial. I got good precision between the two samples. But you said you also had good duplicate precision so that would indicate your method is good.

1 3,090 1,420 117%
2 21,200 12,400 71%
3 43,600 24,900 75%
4 66,400 18,900 251%
5 85,000 69,238 23%
6 99,200 40,781 143%
7 125,000 95,549 31%
8 128,000 107,196 19%
9 164,000 125,603 31%
10 1,080,000 758,760 42%

Bill, your table always shows the higher value of each pair first, so not sure which wells showed decreases and which ones showed increases.