A study focusing on argument as a receiver-contained phenomenon assessed the differential importance of (new) arguments stimulated by a message, as opposed to (old) arguments previously integrated. Data were collected from 207 college students in a pretest, posttest, and delayed posttest design. The subjects provided information on their attitudes toward three topics (sports, job market, landlord-tenant relations), read messages related to the three topics, and completed measures of attitude change. Based on the pretest of student interest in the topics (salience), every subject served as an experimental subject for two topics and a control subject for the third topic. Perhaps the most important discovery was that, on a given topic, people differed in the importance previous beliefs had for them. Belief change was far more likely for people with lightly weighted contexts than for people with firmly anchored views, a finding of clear importance to those interested in either stimulating persuasion or immunizing against it. Persuaded subjects seemed not to have tight contexts surrounding the belief at issue, while resistant subjects showed strong contextual force even prior to the message. Related predictions regarding salience of topic were not supported. The results generally supported a model that characterized the elements of an argument as cognitions quantified by measuring subjective probabilities; the model explained 25% to 50% of the variance in reported adherence to an argument's claim. (RL)