C: any procedure that allows subjects to choose their own experimental conditions

    Correct.

This answer is correct, because a selection bias is anything that disrupts assignment procedures.  Letting subjects choose which condition they are in is a selection bias, because doing so would destroy random assignment.  Random assignment is where each subject in the pool has an equal chance of being selected for a given condition.  If a person chooses her own condition, she no longer has an equal chance of being repersented in each condition.  Recall the Standford prision experiment.  Each subject in the experiment was given an equal chance of being assigned to either the prisoner and gaurd condition.  In this case, selection bias was low.  Thought Question: Imagine if the subjects were allowed to choose their own conditions?  What effect would that have on the results of the experiment?  Would anybody volunteer for the prisoner condition?  What kind of people would volunteer for the gaurd condition?

Thinking about these questions should make you realize how important it is to have a low degree of selection bias.  That is why random assignment is employed.