I just realized my circumference setting is for a 700x32 wheel(shoulda been 700x23). Was my average speed for rides before correction over or under my actual speed, and by how much?

The circumference of the 32mm tire is larger. The wheel sensor measures rotations. A larger circumference/distance value per rotation/time means the computed speed will be higher than actual speed by the ratio of the two circumferences.

Over. The tire circumference is actually smaller so you are getting more revolutions for a set distance. These revolutions are being computed for a larger tire.

BTW: The best way to set the computer is with a roll-out measurement. Inflate the tires to the proper pressure; place the valve stem at the 6 0'clock position; mark the sidewalk with a piece of chalk or something; sit on the seat; and roll one revolution to where the valve stem in back at the 6 o'clock position; then mark it. Next measure the distance between the two marks with a tape measure. Most have CM on them. Enter that into the bike computer. If the computer takes MM just multiply the CMs by 10. Much more accurate.

Yes, your current speed reading is too high.

For the rollout, you should be sitting on the bike as this slightly reduces the circumference...

That's what I was thinking, but I didn't want to face it. So I really didn't average 27.4?

Yeah, that's what I usually do, but I don't usually sit on the bike. I changed the battery in computer, and I guess it reset.

Slightly? Not in my case.

27.4? I think your computer might be set to kilometers.

Ha!

No,it's set for mph, but I only ride 1/2 mile, down this really big hill.

I thought it was off high, but I suck so bad compared to the end of last summer, I thought maybe I had it backwards. Now, I suck worse than I thought.

It's not clear what circumference you are talking about here. The OP is talking about two different tires (with different sizes).

I suppose you mean the effective circumference (while riding) is smaller than the circumference measured for a particular inflated tire.

The "roll-out" method is the best. It's more accurate but not "much more" than the value provided by the cycle computer companies.

The roll-out measurement is actually measuring distance traveled (since there isn't any real circle involved).

A loaded tire/wheel has a flat spot at the bottom (it's not actually a circle; it's some other shape). That makes the distance of the hub axis to the ground shorter (lower). That means the effective radius (and the circumference) is smaller than the radius of the unloaded tire/wheel.

If you could measure the hub-axis height accurately, you wouldn't need to do a roll-out measurement. (For a smooth tire: this distance for a knobby tire will vary during a rotation.)

Note that the difference us also related to the radius. The two radii are approximately "(622+32)/2" and "(622+23)/2". A difference of about 4mm on 361mm or 1.5%.