Fig. 1.

Knowledge level changes after interaction. a Knowledge level was scored in 11 dimensions per session. Differences between interviews of one child (shift, Δ₁₂) and between members of a dyad (gap, Δ_ML) were calculated on a per dimension basis and then averaged. ${\Delta }_{\mathrm{12}}$: knowledge level shift between Interviews 1 and 2, for Children M (${\Delta }_{\mathrm{12}}^M$), L (${\Delta }_{\mathrm{12}}^L$) and control (${\Delta }_{\mathrm{12}}^C$); ${\Delta }_{ML}$: knowledge level gap between Children M and L, in Interviews 1 (${\Delta }_{ML}^{\mathrm{1}}$) and 2 (${\Delta }_{ML}^{\mathrm{2}}$). M1/M2, L1/L2, C1/C2: more knowledgeable (M), less knowledgeable (L), and control (C) children in Interview 1 or 2. b Average knowledge level gap between children in each dyad was significantly smaller after interaction than before. c Shift in knowledge levels. Children L significantly increased their knowledge levels. In contrast, no significant gains or losses were found for either controls or Children M, between whom no significant differences were found either. (*p < 0.05, **p < 0.1, ***p < 0.001, ns: non-significant. Error bars represent SEM, and circles are individual data points)