Depends how long the arm is, I suppose.
Having said that, I decided to look up the thermal coefficient of steel (no idea if this is the material used), and it's 13 x 10^-6 (at 20 °C, but let's go with it for a ballpark figure).
Multiplying 13x10^-6 by a 38 °C shift, by a 2m arm length gave me.... 0.000988m. Pretty much bang on 1 mm. So... unless the thermal expansion of this material is an order of magnitude different, or the arm is 10x the length, it looks like you're on the money.
I went on quite a little journey there, didn't I :D