It’s not who you know: sociometric position did not shape children’s innovation or imitation propensities

Introduction The capacity for cumulative culture depends on the interplay between social learning and innovation, yet the social-ecological factors that shape individual variation in these processes remain poorly understood. Children’s position within peer networks has been theorised to influence learning strategies by shaping exposure to social models and normative conformity pressures, but this prediction remains untested in experimental contexts. Methods We tested whether children’s sociometric position predicted social learning and innovation across three complementary experimental tasks. 256 children aged 7–11 completed sociometric friendship nominations within their school classrooms and three tasks: the Multi-Method Box, a puzzlebox measuring learning strategy choice and innovation-by-modification; a tool innovation task, the hook task; and the Alternate Uses divergent thinking task. We derived three peer-nomination indices — out-degree, reciprocated friendships, and in-degree — and tested their associations with task performance using multilevel models controlling for age and classroom clustering. Results Across all tasks and outcome measures, none of the sociometric indices predicted children’s learning strategy choice, imitation fidelity, innovation-by-modification, tool innovation success, solution method, or divergent thinking performance. In contrast, age consistently predicted innovation-related outcomes: older children produced more reward-exit and tool-entrance innovations on the Multi-Method Box, were more likely to solve the hook task and to use a hook rather than a dragging strategy, and generated more divergent thinking responses. These age effects replicated well-established developmental findings and held independently of children’s network position. Discussion These findings suggest that individual-level dispositions and developmental cognitive gains may carry greater explanatory weight than relational embeddedness for micro-level learning decisions in middle childhood. Network structure may regulate the flow of cultural information across populations without determining the immediate learning strategy choices of individual children in individually administered experimental contexts.