Abstract: The stereotype on scientists is a very important issue in STS, which matters to the maintenance of scientific talent and workforce, risk communication as well as the advance of civic scientific literacy. A consensus in academia is the stereotype on scientists are well established all over the world since Chambers (1983)’s pioneering work. This study however, challenge the methodological and psychological foundation of DAST (draw-a-scientist test) in existence, and proposal a neo-DAST which includes four steps, called “reference drawing”, “comparative coding”, “classified convergence” and “cross verification”. Three groups of subjects was selected from (A) professional masters’ class for public understanding of since in the School of Fine Arts, (B) undergraduate students’ class, and (M) graduate students’ class to examine this new test method (n=71). And it is found that, the stereotype on scientists do exist, but it is also quite doubtable to say that it could definitely be improved by more contact. Dummy regression shows for example, both Group A and Group M generally have less stereotype on scientists, while the former is mainly caused by less contact with science or just because the artists as the reference have stronger stereotype compared with entrepreneurs, and the latter was activated an avoidance mode due to the stereotype threaten. The analysis also implies that, the stereotype of college students sure have some bad effect on choosing science as their further career, where a solution might lie in the enhancement of conterstereotype on scientist by all means.