Well heres what I got, and the variability is well taken.
I wish I could get lots of data points, but this involves moving all my wifes clothes in the master bath closet, pushing out the door to the attic with a pole, putting up a ladder, hoisting myself up and then wiggling on my stomach 17' over to the furnace area over blown in fiberglass insulation
As far as weather, we have this pacific storm system sort of sitting on us. The last three days were pretty close to the same. We had a little wind but nothing special or continuous, just a little gusting. It was in the mid to high 50's overnight and the low to mid 60's during the day. Occasional light rain. While the days were different, they werent materially so.
That having been said, my results were so far apart that we're not splitting hairs over 59 degrees one night and 62 the next.
For almost exactly 24 hours (23:59 by the kill-a-watt), day one of 69 during the day and a set back to 62 at 7pm through 4am, producted .94 Kwh. I think I have this math right, at 533W draw on the furnace fan that means the heat ran for just about two hours total during that one day period; at least the fan did.
For almost exactly 48 hours (47:40...I forgot to go check the furnace at the end of the first day and was half asleep before i remembered, so I decided to make this a 2 day test) of running with the thermostat set to 69 solid, the kill-a-watt registerer 3.35Kwh or about 1.65 Kwh and roughly 3 hours per day of furnace run time, or a 50% higher utilization than with the set back.
Construction wise, my house is standard 3 layer stucco (for those in wood house land, thats a standard 2x4 frame on 16" centers, insulated with 4" of fiberglass. Its tar papered, then a heavy wire mesh is attached and three separate layers of mortar are applied; its a little over 1" thick. The attic is filled with between 9 and 12" of blown-in insulation, depending on the drift of the insulation. It needs a good raking and another 3-4" IMO. The attic has gable vents in 3 places, about 35 soffit vents, and a ridge vent (in other words, excellent attic venting). The house is surrounded on three sides by 16-20' tall shrubs that reduce direct wind access to the walls substantially.
In short, fairly well insulated modern construction. Only foible is the lack of ventilation in the walls behind the master bath tub...and I suspect the guest bathtub as well but I have no access.
So in this decidedly unscientific but close enough for hand grenades test, my nighttime setback produced run times of ~ 2 hours instead of a little more than 3. Based on my gas bill of $156 this month last year when I used roughly the same set-back, thats about a $50-60 a month savings.
Icing on the cake? The first of the two days without the setback, I cooked a 24lb turkey in the electric oven, which had the oven running for about 4.5-5 hours and heated the kitchen up pretty well, so the furnace had some significant help for a portion of the test cycle.
So i'm pretty convinced that the setback helps in a surprisingly substantial way. I sort of had this line of thinking that a short term set back (couple of hours) would be ok, as just the air in the house would cool and quickly be re-heated. I figured a moderate term setback would not work as well, as furniture, walls, floors, cabinets and so forth would all cool after 5-6 hours and the furnace would have to run extensively to re-warm all the solid items in the house. Apparently not so.
What may be a big influence is the moderate temps; we're only looking at 10 or 15 degrees outside vs inside. I may have to rerun this test in a few weeks when this weather system moves out and we start back to getting some low 40's at night and low 50's during the day. At least the weather can be very consistent day to day here, so I should be able to pick a couple of cool ones.
With a 25-30 degree inside to outside delta, the "solid stuff" in the house might cool more quickly and to a lower "core temp" than its doing now with the moderate weather, and cause the furnace to have to work harder and longer to re-warm in the morning, producing more equitable results.
But maybe someone with an actual engineering degree in thermodynamics or some such, or someone who actually went to college, or at least someone who goes to mensa could chime in with an opinion or some other test data.