SS WERE GIVEN VERBAL DESCRIPTIONS OF SEVERAL EMOTIONAL STATES. SOME SS ATTEMPTED TO MATCH EACH DESCRIPTION WITH THE 1 PHOTOGRAPH (OUT OF 11) THAT THE WRITER WAS ATTEMPTING TO DESCRIBE; OTHERS WERE TO RATE EACH DESCRIPTION ON A PLEASANT-UNPLEASANT CONTINUUM, USING EITHER AN 11-POINT SCALE OR A NONNUMERICAL EQUIVALENT. IN ALL 3 TASKS, CONTRAST EFFECTS WERE PRODUCED BY PROVIDING SOME SS WITH MAINLY PLEASANT DESCRIPTIONS, WHILE OTHERS RESPONDED TO PRIMARILY UNPLEASANT DESCRIPTIONS. THESE RESULTS PROBABLY REFLECT CONTEXT-INDUCED CHANGES IN THE EXPERIENCE ELICTED BY THE TEST STIMULI, RATHER THAN CHANGES IN SS' SUBJECTIVE SCALES. SS GIVEN AN UNBIASED SAMPLE OF DESCRIPTIONS WERE MORE ACCURATE IN IDENTIFYING THE "TARGET" PHOTOGRAPHS THAN WERE THOSE ASSIGNED TO BIASED CONTEXTS. (PsycINFO Database Record (c) 2017 APA, all rights reserved)