Maybe a little insight into spin of particles is in order over here.
In QM spin has nothing to do with spinning/rotating but rather has to do with rotational symmetry. For, example take the + sign. In 2d how many degrees do we have to rotate it till it looks exactly like a + sign again? 90 degrees. So QM wise its spin in 360/90 = 4. Similarly, the minus sign has spin 2. Curiously, in 3d you can have topological objects that you have to turn 720 degrees (i.e. twice) for it to look the same again (imagine a twisted belt). These objects will have spin 1/2. The electron (and all spin half particles) are like a twisted belt.
So when we say an electron has spin 1/2 we do not mean that it is rotating like a ball and it has a wee bit angular momentum because of this rotation. But rather its geometry is such that it has to be rotated twice to look the same. Now you can see why its spin will always be there no matter whether it is at absolute 0 or whatever force you apply to it etc. You can never change the geometry of the electron whatever you do.
The direction of the spin (the spin-up/spin-down) thing is basically the two ways of observing the spin (whether you are seeing it as turned once or turned twice).
Now why just geometry turns out to be actual angular momentum in experiments (because you can actually measure electron's angular moment) is a bit tricky to explain. The easiest way to do it is that even in classical mechanics angular momentum - and the conservation of it - is really a result of rotational symmetry (isotropy of space) just like linear momentum - and the conservation of it - is a result of translational symmetry (homogeneity of space). Hence, this behavior just carries on unchanged into QM from classical mechanics.
The relation between spin/geometry and statistics is more difficult to explain. In fact I do not understand it well enough to explain it in simple words.