The Issue with Piano Learning Apps: The Difference Between Scaffolding and Crutches in Learning

Not all help is good for learning. Babies learn to walk on their own, but sometimes with help. It would actually be harmful to give them a literal crutch to help them learn to walk. This is because the ultimate learning goal is to have the child learn to walk without supervision or assistance. On the other hand, we often give training wheels to children learning to ride a bike. These can help a child get a feel for the bike and the motion of pedaling first, before having to worry about balance and the pedaling motion. Once the training wheels are taken off, they then can learn how to balance on the bike while pedaling. Thus, some forms of assistance can be a hindrance to learning (crutches) and others can assist learning (scaffolding).

Unfortunately, in all of the popular music learning apps I have seen, they are utilizing crutches and not scaffolding to teach their users. Thus, users get the illusion they are meeting their learning goals. Most users do not want to learn how to only play the piano while using an app. The general goal of users is to be able to play pieces without an app and to learn generalizable piano skills. The crutch that every app is using is that they all have the displayed notes moving in time concurrently while the user is playing the piece. Thus the user knows visually when they are supposed to press the note. Basically, these apps are glorified Guitar Hero games. Due to some beautiful work from Dr.Wulf, we know that feedback that is used as a crutch hinders the learning of a strategy of how to accomplish a task (e.g. playing the piano) in the absence of feedback (Wulf & Shea, 2002). We can also think about this from a testing effect perspective as well. The testing effect is the simple fact that information that is retrieved from memory has better retention than simply rereading or looking at that bit of information (Gupta et al., 2022). Thus, information that is retrieved from memory will be better remembered than information that is restudied, as in the case of too frequent motor feedback. If feedback is concurrently shown to the learner, they would not have a chance to retrieve it from memory and thus would not benefit from the testing effect. This is the fundamental problem with all piano learning apps that I have seen.

How will MIDIScale address this issue? Although we do not have the exact answer to this problem, we at least know that it exists. One possibility is to slowly take away the crutches from the user. For example, if a user is first learning how to play the C-major scale, they could first practice it with the visual feedback of notes being displayed concurrently on the screen, like the other piano learning apps and while a metronome is playing. After the user has gained proficiency, we can take away the visual feedback of the notes and only keep the metronome. Thus, the user does not become dependent on the visual feedback and can learn to play the scale to a metronome. This can then be extended to taking away the metronome so the user can then play a C-major scale on beat, to any BPM, anywhere, any time.


Wulf, G., & Shea, C. H. (2002). Principles derived from the studies of simple motor skills do not generalize to complex skill learning. Psychonomic Bull Rev, 9(2), 185–211.

Gupta, M. W., Pan, S. C., & Rickard, T. C. (2022). Prior episodic learning and the efficacy of retrieval practice. Memory & Cognition, 50, 722–735.

Join the mailing list to gain access to our bi-weekly newsletter and software updates!

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.