I'd take it back a step further. It'd be getting the right hub. I think the OP was talking about swapping an ABS hub with the bearings installed into it on a non-ABS car. I don't think the races or bearings were swapped.

Recommending a tool to use
 

Hmmmm...don't worry about the accuracy of a device that's used to measure something accurately....that's an idea I'll have to think about.

I'm not against using a dial gauge, but I would like to buy and use one best suited for the task.

There were two separate recommendations for a tool that does not appear to measure to the recommended limit. That's like recommending using a torque wrench that "almost" measures the right torque in place of purchasing the right wrench while simultaneously criticizing someone who uses the "feels tight enough" method. It will get you closer, maybe even close enough, but if a torque wrench that measures on the correct scale is available at a reasonable price, then why not get the right wrench?

If I set the bezel to 0.000, and I get +/- 1/2 of one tick of one increment on a 0.001" tolerance gauge (the hub might move towards and away, right?), then how much metric deflection is there? More or less than 0.01 mm? If less, that's not enough and the bearing would be too tight since the spec is 0.01 to 0.02 mm. If more, then how much more? Is my bearing too tight, too loose or just right?

Using the numbers provided here (not mine, but I'll accept them as accurate mm-inch conversions) and the recommended 0.001" tolerance gauge, I'd have to see at least 1/2 tick (0.012 mm) and not much more than 3/4 of a tick (0.019 mm) of total travel on the gauge or a required accuracy of about 1/4 or 3/8 of a tick to get within the recommended limits.

Why would I recommend a tool that requires me to interpolate to a fraction of the most precise measurement of that tool (aka "looks about right...") when other more accurate tools are available at comparable prices?

Hence my original question: "Is the $20 /set Harbor Freight gauge set-up accurate enough? It almost looks like with this set up you'd need to adjust it to no play whatsoever, or only one-half of one increment on the dial?"

It seems that people who already own a 0.001" gauge say "yeah, it's accurate enough" and I'll accept that since it's already in their tool box. But since I'm considering buying one, maybe I'll get one that lets me use a whole tick, or maybe even two ticks! Nah - that'd be overkill...and I could spend the $3-4 I'd save on a taillight bulb or something.

If you are using a 0.001" gauge, the acceptable range of movement is between a 40% deflection and an 80% deflection, in round figures. A 60% deflection of one increment would be in the middle of the range with a +/- 50% range of error before falling outside the specification.

I use a Starrett indicator calibrated to 0.001 mm. If accuracy and precision are a priority, I highly recommend it.

While you are thinking about, think about how you are going to determine how much force to apply to the hub.

Adjusting the wheel bearings is a lot like drip timing an injection pump: the picture will be a lot clearer after you have done it once.

Problem found. The SE hubs are a little longer due to the tooth ring on the end that the ABS uses. When I was tightening the bearings I was actually tightening the hub into the center of the splash guard. As it was grinding it down the bearings would become loose. I should have taken a picture but my camera was at work. Needless to say, I switched hubs and everything is good.

If you are going to change hubs like that, you should always measure clearances. Probably should have grabbed the steering knuckles and backing plates as well.