Moving Time and Elevation Data Wildly Inconsistent Between Garmin Devices

What's causing this discrepancy, and is there a way to improve the accuracy of moving time?

Could be a number of factors. Among others the hardware, the software, the settings, and the exact position of the device. As for the settings, make sure both devices use the same recording mode (if possible). Typically, the "Smart Recording", gets the coordinates from GPS when some of the parameters (speed, direction, HR, elevation,...) change, so each device may get the keypoints at different places. The "Every Second" setting is more accurate, but at a 6hr hike it may generate a huge file, and drain the battery more than the Smart Recording. The worst at all is the UltraTrac mode - it is practically useless, because it takes the position each minute only, and the accuracy is then quite miserable if the condition are not ideal. I would definitely avoid it. You can also enable the option GPS+Galileo on both devices, in the settings of the concerned Activity profile. It helps getting the lock on more satellites quicker, when under difficult conditions. It will also drain the battery a bit more, but the difference is not significant.

On a similar note, Garmin Connect's pace chart doesn't show accurate data when pace is slow.

Yes, there is a lower limit of speed that is still considered "hiking". I do not think you can adjust the limit. You can try other activity profiles, such as the Walk, or creating a Custom activity profile, to see whether it changes there.

On a similar note, Garmin Connect's pace chart doesn't show accurate data when pace is slow. It only shows as 1:00:00 min/km, like this: https://preview.redd.it/tvzxrg9dii261.png?width=1040&format=png&auto=webp&s=752654495768f8c8411a91755a60d46af8ba426a The slow pace of 1:00:00 min/km, which is also off from the actual pace, corresponds to non-moving time. Again, how do I get accurate data for a slow pace activity such as hiking?

As useful as pace in min/km or min/mi can be for running (although some would disagree), the problem with this metric is that there's no actual lower limit (i.e. standing still would be ∞ min/mile or /km), so I think they have to pick a cutoff value. And the choice of cutoff changes the scale of the graph, of course. But they have to pick an arbitrary cutoff *somewhere*. There's probably better ways to do it than it's currently implemented (like analyzing the data and seeing what the typical pace is like.)

It could also just be that you're hiking slower than Garmin's pace threshold for "moving".

Thank you all for your input. Just FYI, my Fenix 6 was in GPS+Galileo mode, and recording every second. GPSMap 66 was in GPS only mode, and set to 'normal' record interval. I found that when conditions were less than favorable (under tree cover, etc.), GPSMap 66 produced more accurate GPS track, even though it was recording less frequently and had only GPS enabled. I guess it's due to the GPS chipset difference? Btw, would enabling WAAS/EGNOS help obtain more accurate moving time data for GPSMap 66?

GPSMap 66 produced more accurate GPS track, even though it was recording less frequently and had only GPS enabled. I guess it's due to the GPS chipset difference?

You do not expect the same accuracy from a extremely miniaturized watch as from a specialized handheld GPS device, or do you really? It would be sad if the handheld device could not get a more accurate position. The chipset is not the only factor, the antenna, and the remaining electronics play the principal role in it.

would enabling WAAS/EGNOS help obtain more accurate moving time data for GPSMap 66?

That's what WAAS/EGNOS was designed for, so if it is availble for your device, then it certainly should.