Determinants of experimental neurosis in cats

In order to assess the relative effectiveness of Masserman's contingent method and Wolpe's noncontingent method of inducing experimental neurosis in cats, cats in four independent groups were trained to a uniform criterion to eat in a standard apparatus and then shocked either contingently upon eating or noncontingently upon eating or noncontingently. Cats shocked noncontingently were matched with those shocked contingently for the number of shocks administered. Ss then were assessed for persistence and/or generalization of the inhibition of eating in the same or in a discriminably-different setting. Masserman's contingent method induced significantly greater inhibition at the 0.5 level. This finding is consistent with prediction that would be made on the basis of the delay-of-reinforcement gradient. It is suggested that the contrasting results obtained by Wolpe and Smart may be attributable to confounding variables introduced unintentionally.