The effect of zero-inflation in the data was investigated via ARL comparisons between the various control schemes under the different
models. More specifically, we compared the PINAR(1) and ZIPINAR(1) (resp. PINARCH(1) and ZIPINARCH(1)) Shewhart and CUSUM
charts in the detection of shifts in process mean 0. The results showed that the excess in zero counts requires a proper adjustment for
the values of the design parameters of each scheme in order to preserve the false alarm rates to the desired levels. Moreover, as the percentage
of zero counts increases, the performance of the ZIPINAR(1) Shewhart chart improves, while the performance of the ZIPINAR(1)
CUSUM chart becomes worse. For the case of ZIPINARCH(1) process, the analysis showed that the respective Shewhart and CUSUM
charts are much more affected (especially the latter) by the presence of zero-inflation compared with their ZIPINAR(1) counterparts.