In this doctoral thesis we study stealth CMEs: solar coronal mass ejections that are clearly observed in coronagraph data but do not show significant low-coronal or on-disk signatures of eruption. This lack of coronal signatures makes it challenging to determine their source region and predict their trajectory throughout interplanetary space. We identify 40 such events and investigate their properties both observationally and statistically. We find that our sample size is insufficient to determine the scaling law for the CME angular width reliably. We therefore analyze in general what the effect is of a limited sample size on the estimation of a power law parameter. Armed with this knowledge, we return to our sample of stealth CMEs, re-analyze the power law for their angular widths and compare the results to the power law found for normal CMEs.