In a repeated shifts experiment four independent groups of thirsty rats received the following treatments: LSLS, LLLS, SSLS, and SSSS, with each letter denoting the magnitude (large or small) of sucrose reward received in each of the four phases of the experiment. While no negative contrast effect (NCE) was obtained in Phase 2, a very reliable positive contrast effect (PCE) was found in Phase 3. Moreover, a significant NCE was obtained in Phase 4. The results were explained in terms of the relative rather than absolute effects of reinforcement. Crespi (1942) was probably the first investigator to popularize the phenomenon of incentive contrast. He reported obtaining both a positive PCE) and a negative (NCE) contrast effect. Subjects shifted from small to larger rewards were observed to run at a faster rate than would be expected from a control group kept on large reward all the time (elation effect or PCE). Conversely, when subjects were shifted from large to a smaller reward, they undershot the level of the subjects receiving small reward all the time (depression effect or NeE). Attempts to replicate Crespi's findings of PCE have failed repeatedly. When appropriate controls for the ceiling effects were introduced, PCE was consistently obtained (Shanab, Birnbaum, & Cavallaro,' 1974). Recently, however, some studies have obtained PCE without such controls (cf. Weinstock, 1971).Crespi's findings of NCE in the runway based on shifts in food reward have been successfully replicated in many laboratories (Dunham, 1968). However, when either magnitude or concentration of liquid sucrose (Goodrich, 1962;Homzie & Ross, 1962;Ison & Rosen, 1968;Rosen, 1966;Rosen & Ison, 1965), or both magnitude and concentration of liquid sucrose were simultaneously downshifted (Barnes & Tombaugh, 1973;Flaherty, Riley, & Spear, 1973;Ison & Rosen, 1968), no NCE was obtained. The dependent variable in the previous studies was response speed. It should be noted that a NCE was reported with liquid sucrose when the dependent variable was either the frequency of the licking response (Vogel, Mikulka, & Spear, 1968), or the barpress response (CoIlier & Marx, 1959; Weinstein, 1970a, b).A closer examination of Crespi's design for positive contrast reveals that he used a double-shift in that his peE subjects originally recieved large reward followed by a shift first to small reward and then back to large reward. Crespi (1942) reasoned that in order to obtain a Reprint requests should be sent to Mitri E. Shanab,

No abstract

In Experiment I, hungry rats first received 10 pellets of either 4% or 16% sucrose, and later received one pellet of the same concentration as before. Control subjects received one pellet of either 16% or 4% throughout. A significant negative contrast effect was obtained with both concentr~tions. In Experiment II, three groups received four pellets of either 4%,8%, or 16% ~u~rose concentr~tlon before receiving the same number of pellets of 8% concentration. Neither a posItive nor a negative. contrast effect was found. The results were discussed in terms of both the frustration and generalization-decrement hypotheses.

In Experiment I, thirsty rats first received 1.50 cc of either 4% or 16% sucrose solution and l~ter received .25 cc of the same solution as before. Control subjects received .25 cc of the appropnate solution throughout. In Experiment II, experimental subjects were shifted from 1. 75 cc of 8% to .20 cc of the same solution, but all subjects were later given 1.75 cc sucrose. reward. In neither experiment 'Yas a reliable negative contrast effect obtained. However, in Experiment II a significant positive-contrast-hke effect was obtained. The results were discussed in terms of the older theories of Crespi and Helson as well as the new theory of Capaldi, all three of which emphasize a relative rather than an absolute view of reinforcement.