This study explores American and Greek primary pupils’ visual images of scientists by means of two nonverbal data collection tasks to identify possible convergences and divergences. Specifically, it aims to investigate whether their images of scientists vary according to the data collection instrument used and to gender. To this end, 91 third-grade American and Greek pupils were examined. Data collection was conducted through a drawing task based on Chambers ‘Draw-A-Scientist-Test’ and a picture selection task during which the children selected between (...) 14 pairs of illustrations those that were most probable to represent scientists. Analysis focused on stereotype indicators related with scientists’ appearance and work setting. Results showed that the two groups’ performance varied significantly across the tasks used to explore their stereotypic perceptions, although the overall stereotypy was not differentiated according to participants’ ethnic group. Moreover, boys were found to use more stereotypic indicators than girls, while the picture selection task elicited more stereotypic responses than the drawing task. In general, data collected by the two instruments revealed convergences and divergences concerning the stereotypic indicators preferred. Similarities and differences between national groups point to the influence of a globalized popular culture on the one hand and of the different sociocultural contexts underlying science curricula and their implementation on the other. Implications for science education are discussed. (shrink)