Human versus computer responding - Simulating stimulus equivalence experiments using Enhanced Equivalence Projective Simulation
MetadataShow full item record
The purpose of this study was to explore if it is possible to simulate participant responding in a stimulus equivalence experiment, using the enhanced equivalence projective simulation model from Mofrad et al. (2021). Five human participants and 45 simulated participants, characterized by different model parameter values, were exposed to the same matching to sample procedure. In both experiments, the participants in average did not attain the criterion of 90% correct trials during equivalence test. The mean human participant equivalence test score was 84%, while the mean simulated equivalence test score was 82%. Human participants passed the baseline and symmetry tests with a mean score of 97%, against a mean score among simulated participants of 89%. The simulated participants, when using the same parameter settings, were homogenous in terms of their response patterns, while the human participants exhibited a larger variation. A suggestion for further work, groups of simulated participants with different sets of parameter values, could perhaps yield inter-group responding similar to the observed responding in human experiments. A practical application of these findings is to explore how simulations and human experiments can be compared at a larger scale. A possible solution could be to add the simulation code to the MTS software that is used in the human experiments, so that simulations are automatically run and compared to human responding.