In a 2 × 2 design, 85 subjects were asked to estimate the size of angles (direct influence) that were either 90 or 85°, after being confronted with incorrect judgements of a majority (88 per cent) or a minority (12 per cent) of people estimating the angles at 50°. Additionally, pre- and post-test measures were used to establish indirect influence on subjects' judgements pertaining to acute angles (i.e. on the estimation of the length of lines constituting the angles, and on the imaginary weight of figures represented by these angles). Overall, little direct influence is observed. This may partly be due to the introduction of a denial of the credibility of the source in all conditions. In fact, some evidence of direct influence is only found in the majority–85° angles condition. An instance of indirect influence (on the estimation of length of lines) appeared as the result of a majority stance when the angles in the experimental phase were 90°. When these angles were 85°, indirect minority influence (on the estimation of weight of figures) was observed. These effects had been predicted on the basis of the hypothesis stating that indirect majority influence would be possible when subjects expected consensus on the correct response (in the 90° angles condition), without being able to reach consensus at the manifest level (because of the denial and the restriction imposed by the clear shape of 90° angles). Indirect minority influence was hypothesized to be stronger in a situation that allows for diverse responses (i.e. for 85° angles).