There is a phase shift associated with reflection which may affect how the s and p components are combined in the reflected beam vs the incident beam. However that phase shift depends also on whether the boundary is one of increasing vs decreasing index of refraction. You'll have to study up on these issues and see if your reasoning still holds.
For example if, IF it is valid to view the reflected beam as the mirror image of what the incident beam would look like if it continued unreflected, then it would be the mirror image of a Left circular polarized beam and thence be a Right circular polarized beam (or vice versa). But is that a valid analysis? I think this question is intended to have you explore these issues. I haven't thought through the question fully myself, but rather am bringing up issues I would double check before answering. Some may be irrelevant in the final analysis. So don't assume I'm hinting at a specific answer, yours may be right, I just don't think you argument is quite sufficient to be sure.