Effects of serial and concurrent training on acquisition and generalization.

Five preschoolers with developmental delay learned letter names and sounds. The team compared two ways to show the letters. Serial meant one letter at a time. Concurrent meant three letters shown together. Each child tried both ways in an alternating-treatments design. Sessions looked like typical discrete-trial drills: model, response, praise, next trial.

Concurrent presentation won. Kids reached mastery in fewer trials. They also named or sounded untrained letters better at the end. The serial way took longer and generalization was weaker.

O'Neill et al. (2022) found the same "more-is-faster" pattern. Progressive prompt delay beat fixed 2-s or 5-s delays for expressive labels. Both studies show that moving, rich input beats static, thin input.

Koegel et al. (2014) tested receptive labels in autism. They skipped simple discriminations and started with a 3-item conditional array. That tweak, like concurrent letters, piles stimuli together and cuts learning time.

Giunta-Fede et al. (2016) compared continuous versus discontinuous data collection during tact training. Continuous tracking caught learning gains sooner, but the teaching itself stayed serial. The 2014 letter study goes one step further: it changes the teaching setup, not just the data system.

Stop drilling one target at a time. Put 2-3 stimuli on the table from the start. You will save trials and see broader generalization. This is an easy Monday switch: swap your single-card array for a small field. It works for letters, sounds, and possibly other discrete skills.