In the top 5% of all research outputs scored by Altmetric

Hi all, with the great Roddy Grieves we have a new paper out: a meta-analysis showing that the Sunburst Maze experiment of Tolman, Richie & Kalish (1946), often used as one of the pillars of the Cognitive Map theory, hasn't really been replicated despite many attempts. Basically, the original paper shows that rats preferred a "shortcut" path to a reward location and Tolman later argues that this (together with other results) supports the idea that rats have an internal "cognitive map". BUT most replication attempts instead find that rats or humans prefer paths adjacent to their training route, outer paths, cued paths (humans only) or unremarkable paths. A large chunck of the rat studies showed no path preference at all... And this is probably just the top of the iceberg: the studies that have been published (cf. publication bias for "positive" results)! The title says it all I think:  Tolman's Sunburst Maze 80 Years on: A Meta-Analysis Reveals Poor Replicability and Little Evidence for Shortcutting Link: https://lnkd.in/dQreKmRb Conclusions: the Sunburst experiment is not a good test of shortcutting and does not actually demonstrate that animals (rats, humans or marmosets) can do map-based shortcutting.  What would be a better demonstration of map-based shortcutting? We need to think about that. Does this destroy the cognitive map theory? Not at all - several independent lines of evidence support that theory. Still, it confirms that single papers with shiny findings are rarely enough to establish a scientific fact, and that as researchers it is crucial to give more importance to replications and "negative findings". Check our thread on Mastodon (https://lnkd.in/dDJB57Zb) or Bluesky (https://lnkd.in/dRXs_zwB) and feel free to post any questions or comments!

In the top 5% of all research outputs scored by Altmetric

Hi all, with the great Roddy Grieves we have a new paper out: a meta-analysis showing that the Sunburst Maze experiment of Tolman, Richie & Kalish (1946), often used as one of the pillars of the Cognitive Map theory, hasn't really been replicated despite many attempts. Basically, the original paper shows that rats preferred a "shortcut" path to a reward location and Tolman later argues that this (together with other results) supports the idea that rats have an internal "cognitive map". BUT most replication attempts instead find that rats or humans prefer paths adjacent to their training route, outer paths, cued paths (humans only) or unremarkable paths. A large chunck of the rat studies showed no path preference at all... And this is probably just the top of the iceberg: the studies that have been published (cf. publication bias for "positive" results)! The title says it all I think:  Tolman's Sunburst Maze 80 Years on: A Meta-Analysis Reveals Poor Replicability and Little Evidence for Shortcutting Link: https://lnkd.in/dQreKmRb Conclusions: the Sunburst experiment is not a good test of shortcutting and does not actually demonstrate that animals (rats, humans or marmosets) can do map-based shortcutting.  What would be a better demonstration of map-based shortcutting? We need to think about that. Does this destroy the cognitive map theory? Not at all - several independent lines of evidence support that theory. Still, it confirms that single papers with shiny findings are rarely enough to establish a scientific fact, and that as researchers it is crucial to give more importance to replications and "negative findings". Check our thread on Mastodon (https://lnkd.in/dDJB57Zb) or Bluesky (https://lnkd.in/dRXs_zwB) and feel free to post any questions or comments!