We have noted previously that the control chart is statistically equivalent to testing a hypothesis. In fact, it is possible to use various statistical-hypothesis testing procedures as alternatives to control charts. For example, contingency tables can be used instead of the fraction nonconforming control chart to study a set of preliminary samples for lack of control. The contingency table can also be used to study defects per unit in preliminary samples, instead of c charts or u charts. Duncan (1986) has pointed out that there is little difference in the two approaches, except that it is customary to compute the OC curve for a contingency table in terms of an average difference from the nominal value of the parameter in question (say p ), whereas in the control chart analysis, we usually tabulate the OC curve in terms of a single sample having a value of the parameter that exceeds the nominal value.
The analysis of variance (ANOVA) may be substituted for the usual xbar and R charts when examining preliminary samples for lack of control. However, the two procedures do not always produce identical results. ANOVA is a tabular format and is difficult to evaluate the out-of-control conditions.