Compared to efficient decoders (who excel on tests of nonverbal decoding ability), inefficient decoders may (1) use a restricted repertoire of nonverbal cues; (2) use the same repertoire but be less sensitive to the cues; or (3) be more sensitive to cues of high informative utility and less sensitive to cues of lesser utility. Subjects were asked to identify romantic heterosexual couples in photographs. Potential cues were generated by think aloud protocols (Study 1); their presence and informative utility were determined through ratings (Study 2). Study 3 estimated cues' influence on "couple" decisions. Decisions were affected by all the informative cues ("proximity," "contact," and "context") with efficient decoders being more sensitive to "proximity" than inefficient decoders. However, inefficient decoders were more sensitive than efficient decoders to uninformative cues suggesting couples' similarity. Hence, the present test does not measure sensitivity (Hypotheses 1 and 2) but measures sensitivity-tuning (Hypothesis 3).