Cosmic Shear from STIS pure parallels III. Analysis of Cycle 9 pure parallels

Following the detection of a cosmic shear signal at the 30 �� scale using archival parallel data from the STIS CCD camera onboard HST in the previous paper of this series, we analyzed a larger data set obtained from an HST GO pure parallel program. Although this data set is considerably larger than the one analyzed previously, we do not obtain a significant detection of the cosmic shear signal. The most likely cause of this null result is that the signal is comparable to (or less than) the random measurement of the noise. We also show that multiple systematics plague the STIS CCD data for its use in cosmic shear programs, and in particular the continuous degradation of the CCD charge transfer efficiency after 4 years in space, which is a source of additional noise that is still difficult to correct for in the absence of physical models.