Hm... I always check my gage on st st. Though it would be curious to find out how the pattern will change that; but as far as I know, gage is generally done off st st unless the pattern states othewise (sometimes it will).
And as for doing a gage check on straits or in the round, here's two schools of thought.
1. Doing it on strait needles: When you knit back and forth on strait needles, your gage may be off a bit from working in the round because knitters tend to use different tension between knit and purl rows. However, it can give you a good aproximation of what your gage will be. It really shouldn't matter whether you do it on dpns or not, except that it's certainly possible that the ACTUAL diameter of the needles may be slightly different between brands/styles of needles. All size 5 needles are not exactly the same diameter, though they are probably really close. This is due to manufaturing technique etc. I'd work on the needles you were going to knit you socks in.
2. Doing it "in the round" on strait needles: This is a strange technique that many swear by. It allows you to check your gage using only knit stitches, just like you would create working in the round. The explanation is a little wierd.
CO a number of stitches on dpns (or circular needles, as it works the same for them), do a garger edge if you want. Then knit the first row of the gage test section. Instead of turning the work (you will in fact not turn the work again), slide the stitches over to the other end of the needles. The working yarn, which would normally be coming from the first stitch on the end of the needle, will now be coming from the last stitch on the needle. Loop the feed yarn around the back of the fabric to bring it around to the first stitch and begin to knit with it. Knit across the row. When you get to the end, slide the stitches to the other end of the needle, loop the yarn across the back, and begin to knit again. Repeat.
What this will do is create a swatch of st st with lots of strands across the back. You can snip the strands so they are not loops so the swatch will more comfortably lay flat when you measure your gage. Let me know if this doesn't make sense, and I will knit up a test swatch, take some pictures, and show you visually.
I guess the "proper" way of testing a swatch for in the round knitting is using method 2. Most books will tell you to do it that way.
However, for what it's worth, I use method 1. I've found that my gage is pretty darn close working flat and working in the round. Granted, that's taking into account my personal knitting style and it may not work for you.