The present study set out to assess if the different probabilities reported in the literature of obtaining equivalence after baseline training with MTO and OTM protocols could be attributed to individual differences and, if so, whether equivalence formation could be facilitated by using familiar stimuli as nodes. In Experiment 1, 16 preschool children were trained on four sets of 2-choice match-to-sample tasks, eight with a OTM protocol (A–B, A–C, A–D, A–E) and eight with a MTO protocol (B–A, C–A, D–A, E–A). For four OTM and four MTO children only abstract stimuli were used. The other four OTM children and four MTO children received the same training but with familiar stimuli as nodes. All children received tests for equivalence (first) and symmetry (second). In Experiment 2, eight children who served in Experiment 1 participated again, four who had passed the equivalence test, and four who had failed that test. All children received the same baseline training as in Experiment 1 but with the opposite type of nodes (abstract instead of familiar, and vice versa) and training protocol (MTO instead of OTM, and vice versa). The results showed that (a) the children's performances (training and testing) were not affected by the training protocol; (b) equivalence formation occurred more readily when being trained with all abstract stimuli than when familiar stimuli served as nodes; and (c) most children who passed or did not pass the equivalence test in Experiment 1 repeated their performance in Experiment 2, irrespective of the conditions that were used.