This research supported the hypothesis that hypnosis can be thought of as a set of potentially modifiable social-cognitive skills and attitudes. A low-interpersonal-training treatment devised by Gorassini and Spanos (1986) was compared with a treatment designed to modify not only cognitive factors but also to augment rapport with the trainer and diminish resistance to responding (high-interpersonal training). Fifty percent of the initially unhypnotizable subjects in the high-interpersonal condition tested as being highly susceptible to hypnosis (high susceptibles) at posttest on the Harvard Group Scale of Hypnotic Susceptibility (Shor & Orne, 1962); 25% of the unhypnotizable subjects in the low-interpersonal condition responded comparably. Eighty-three percent of the medium-susceptibility (medium susceptibles) subjects tested as being highly susceptible at posttest in both conditions. Practice-alone control subjects' performance was stable across testings. The study was the first to demonstrate that treatment gains generalize to a battery of novel, demanding suggestions (generalization index) that have been found to differentiate highly susceptible subjects from unhypnotizable simulating subjects. The importance of rapport was evidenced by the finding that rapport ratings paralleled group differences in hypnotic responding and that rapport correlated substantially with susceptibility scores at posttest and with the generalization index. Whereas initial hypnotizability scores correlated significantly with retest susceptibility scores, initial hypnotizability failed to correlate significantly with the generalization index.