The present study compares the effects of two intervention procedures on children’s ability to tell coherent stories, taking into account the causes of events and the internal states of the characters. Using an image-based story, one procedure investigated whether a verbal interaction between the child and a less competent peer (simulated by a hand puppet) (the PINT procedure) or a conversation focusing children’s attention on the causes of the events (the CONV procedure), could promote more coherent and evaluative narratives in 6- to 8-year-old French-speaking children. After the first monological narrative produced after having seen the images (NS1), the children produced a second narrative after one or the other intervention (NS2), and one week later, another narrative for the same story, to assess the stability of the possible gains obtained immediately after the procedure, and a narrative for an analogous story to assess whether improvements were generalizable. During the PINT intervention children told the peer-puppet two ‘interactive’ narratives (INTNS1 and INTNS2). Results confirm the positive effects of the CONV procedures obtained in previous studies. The PINT procedure used here did not produce the expected beneficial effects. It produced instead a ‘depressor’ effect on the narrative produced immediately after the procedure compared to the initial one. The results are explained in terms of the specific characteristics of the PINT procedure designed in this study.