The triggering mechanism(s) for the substorm expansion phase onset is one of the outstanding issues in magnetospheric physics. Previous studies have shown that an impingement of the high solar wind dynamic pressure may lead to the expansion phase onset of a substorm. These previous studies typically used a negative magnetic bay as the main proxy for a substorm, but we now know that some magnetic bays are associated not with substorms but with the enhancement of convection. Therefore it is reasonable to cast doubts on the compression trigger mechanism. In this study we use classical substorm onsets, auroral breakups, which are the most reliable substorm onset indicator when identified with global auroral images, to reinvestigate this issue. We examine 43 interplanetary shock events that occurred between 1996 and 1999 with simultaneous global auroral images from the Polar ultraviolet imager images and the auroral electrojet indices. It is found, indeed, that ~52% of the shocks produce magnetic bays (AL < -100 nT). While most of the shocks enhance auroral luminosities, only 4 events (~9%) appear to have an auroral breakup preceded by an SSC/SI by 20 min (two events by 10 min). These results strongly indicate interplanetary shocks can produce negative magnetic bays but not auroral breakups (thus christened compression bays). An examination of precisely timed interplanetary magnetic field (IMF) during 11 substorms that occurred near the shock-induced SSCs/SIs indicates that northward turnings of IMF occurred more than 50% of time. Given such a low probability, it is concluded that shock compression is not likely to trigger substorms but enhances magnetospheric currents and auroral particle precipitation. In line with many previous study results, on the other hand, the northward turnings of IMF can be a plausible substorm triggering mechanism.