Given two identical samples of heat treatable steel in both composition and geometry, if one of the pieces is annealed or hardened would there be a measurable difference in their thermal expansion coefficient?
In my experience, tables/data listing the thermal expansion coefficients of different materials do not make any differentiation between hardened/annealed etc. states of any given material. The degree of accuracy to which the TEC can be measured is limited by the tolerances to which each material is made up of, and this variability will likely be larger than any change from heat treatment.
One related thing to note, is that when you heat treat one of your samples, it will likely change size as a result of the treatment. See https://vacaero.com/information-resources/the-heat-treat-doctor/1316-dimensional-changes-after-heat-treatment.html for more details
I don't see how they can have the same value. Hardening or annealing affects the crystal structure of material and in the case of steel, how much of the carbon (and other hardening elements) gets absorbed into the iron. Both of these will affect how they react to changes in temperature. Is the coefficient of thermal expansion higher or lower? That part, I can not tell you. I have searched and searched on the web looking for examples where the hardening is listed with the coefficient of thermal expansion and I came up empty. I expect that the variability due to differences in steel chemistry has more effect than that of hardening. (Keep in mind that, say, 1040 steel can have between 0.370 and 0.440 % carbon and between 0.60 and 0.90 % manganese.)
In a career as a metallurgist , I never saw a distinction noted for different thermal expansion rates. However in a hardened condition steel will be expanded slightly. And for some high alloys that may be ferritic or austenitic ( depending on heat-treatment) at ambient temperatures, there will be a coefficient difference.