Kuleshov’s montage experiments have arguably been a key impetus for inauguration of film theory. Yet, although cognitivists – and even some continental film philosophers – have long appreciated the importance of neurological and psychological studies for understanding film, they rarely undertake experiments themselves. Instead, the work is primarily done by psychologists with special interest in film. This paper advocates for a deeper engagement with the experimental method in film studies, through design and/or criticism of specific experiments. First, to dispel the longstanding disciplinary skepticism against the method, I propose that arguments against cognitivism as methodologically imperialistic conflate the methods of analytic philosophy and scientific experiment. I then retort to strong (D.N. Rodowick) and moderate skepticism (Malcolm Turvey) about the experimental method. Against the former I argue that 1) some questions in film studies demand experimental answers, and 2) these experiments do not transform film studies into a science of film, and 3) inferences drawn from experiments are not incommensurable with humanistic inquiry. In the latter case I point out that although there is a difference between humanistic and natural phenomena and the principles behind them, there are some principles behind humanistic phenomena which are discoverable through experimental method. Second, to illustrate the importance of the experimental method I draw attention to the fact that a key assumption in film studies – that fiction films change our beliefs about the actual world – is an empirical claim still awaiting experimental proof. I specify how one experiment (co-developed with Ed Tan) testing this assumption might look. I also pay special attention to problems of replicability and representativeness at the crux of the current crisis in psychology. In conclusion, I invite film scholars to a close reading of the proposed experimental design as a way of coming to grips with challenges, opportunities, and the potential blind spots of experimental work.