My relevant experience extends only to vibration tests on various mechanical enclosures for electronic equipment and the like. A temperature change of ~ 10C by itself is typically responsible for expansion of less than one part in a thousand or less, so no much effect should be seen there. But what did seem to happen is that some separate, but adjacent, rigid components in the structure, with their own resonances, would apparently via expansion compress each other and act as one rigid structure, and thus have a quite different resonance, up to an amplitude that re-disconnected the components. At least we would see different resonances exposed, and that was a hypothesis.
That hypothesis tends to match my observation of varying house resonances over temperature/humidity from aircraft going overhead, etc, and I've tended to chalk it up to wood frames, glass, drywall closing up, or opening, gaps. So for the house: all hypothesis, no test data.