Purpose The gathering of human intelligence (HUMINT) is of utmost importance, yet the scientific literature is silent with respect to the effectiveness of different information elicitation techniques. Our aim was to remedy this by conducting the first scientific test of the so‐called Scharff technique (named after the successful German WWII interrogator). Method We developed a new experimental paradigm, mirroring some main features of a typical HUMINT situation. The participants (N = 93) were given information on a planned terrorist attack, and were instructed to strike a balance between not revealing too much or too little information in an upcoming interview. One third was interviewed with the Scharff technique (conceptualized to include four different tactics), one‐third was asked open questions only, and the final third was asked specific questions only. The effectiveness of the three techniques was assessed by a novel set of objective and subjective measures. Results Our main findings show that (1) the three techniques did not differ with respect to the objective amount of new information gathered; (2) the participants in the Scharff condition perceived (as predicted) that it was more difficult to read the interviewer's information objectives; and (3) the participants in the Scharff‐ and the Open‐question condition (incorrectly) perceived to have revealed significantly less information than the participants in the Specific question condition. Conclusions We presented a new experimental paradigm, and new dependent measures, for studying the effectiveness of different information elicitation techniques. We consider the outcome for the Scharff technique as rather promising, but future refinements are needed.