A colleague of mine claims the multiturn ones are much less stable with temperature, based on testing with his previous employer. The datasheet accurately represents the stability of the resistive element, but thermal expansion of the mechanics apparently causes small jumps in resistance with temperature, the wiper physically moves. These jumps, so he tells me, are discrete steps, and have some hysteresis, so are hard to compensate for.
I have not tested it myself, but it certainly seems plausible, and as a result I tend to use coarse/fine pots rather than multiturns in precision designs.