What happens in the arrangement is that you are applying a force away from the center of gravity.
Imagine you have a pencil on a flat surface and you try to push. If you don't push it around the center of gravity you will cause it to rotate.
Although using linear bearings (especially longer ones) will mitigate the problem, you will be creating more normal forces, therefore more friction, and the motor will draw more Amps in order to provide the required torque. And that would be bad design.
Surprisingly, even if the actuator is in the middle you might run into problems depending on the tolerance of the bearings. That was an experience I had when designing a (budget) linear translation stage (that was almost 10 years ago)

10 or so were build from these, (with different lengths - so only the rods had different lengths). Generally they worked well, however the initial prototype was troublesome with respect to getting the alignment of the bars correct, and using linear bearings of proper tolerances (too tight tolerances, were not good when there were misalignments in the rods, while too loose were making the stage move like a drunk sailor in an alley).
Generally, the most important thing is: not to overconstrain the assembly. And that I'm afraid is something you can't read, you have to make mistakes and learn.
And just to be clear: what I said in the previous lines, does not mean that the power screw in the middle is worst that two power screws either way. The dual motor design if not properly implemented (e.g. if you don't make sure that both motors are sychronised), can end up is similar or worse self locking problems.