The effect of errorless versus errorful learning on generalized motor program learning and parameterization learning

Abstract

Objectives: A contradiction exists between the mechanisms associated with errorless learning (EL) and those attributed to contextual interference (CI). EL discourages conscious attempts to modify technique (Maxwell, Masters, Kerr & Weedon, 2001), while the CI literature suggests benefits of practice structures that encourage conscious engagement in (trial-to-trial) movement variability (e.g., variable practice). This paradox was explored by applying measures reported in the CI literature to an EL paradigm. The main goal was to dissociate variability in Generalized Motor Programs (GMPs) and parameterization in EL. Design and Method University students learned a darts-like ball throwing task over 300 trials under either errorless or errorful conditions and performed 150 retention and transfer trials. In a learning session, both groups threw golf balls towards a target at 4 m distance for 6 blocks of 50 trials. The errorless group started with a target size of 95x95 cm. After each block of trials the dimensions of the target were decreased by 15 cm. The erroful group threw at smallest target of 20 x 20 cm exclusively. More than one day after the learning session, participants performed a total of 150 (3 blocks of 50) trials to targets at distances of 3.5, 4 and 4.5 m respectively. During the test session only the smallest target size was used. Trial-to-trial GMP variability was quantified as the Euclidean distance of the time- and amplitude normalized kinematic data (Jaitner, Mendoza & Schöllhorn, 2001). In line with Wulf and Schmidt (1994) the scaling factors of time and amplitude represented parameter settings. Results and Conclusion The errorless group committed fewer errors during both learning and test sessions. Preliminary analyses showed that the errorless learning group combined low GMP variability with high variability in parameterization. It is concluded that errorless learning yields more effective generalized motor program learning and parameterization learning.