Chiral molecules act as strong spin filters for transmitted electrons (chiral-induced spin selectivity). The interplay of geometry and spin mediated by spin-orbit coupling is commonly assumed as the cause of the effect, but the theoretical description remains incomplete. In this thesis, two models for electron transport through helical molecules were investigated: an atomistic tight binding model for the molecule helicene and a simple continuum model for an electron in a helix-shaped potential. In an attempt to cover the middle ground between phenomenological tight binding approaches and detailed first principle simulations, the helicene model starts with a lattice of carbon atoms represented by a minimal basis of local atomic s- and p-orbitals including electronic nearest-neighbor and spin-orbit interactions. Löwdin partitioning is used to reduce the model to a p-orbital tight binding representation, providing numeric values for all the couplings dependent on geometry. Transport calculations showed helicity dependent spin polarization several orders of magnitude smaller than experimentally observed. To understand the effect on a more fundamental level, an electron moving through a helix-shaped confinement potential in 3D space with spin-orbit coupling was considered. By taking the limit of strong confinement, an approximate model with one-dimensional configuration space (the helix) was obtained. Novel onsite spin-orbit coupling terms appear in the effective Hamiltonian, leading to sizeable spin polarization in transport calculations. These new terms are thoroughly justified by the adiabatic limiting procedure which was adapted to include spin-orbit coupling and might thus provide one of the missing pieces for the theory of chiral-induced spin selectivity.