The design story behind a Hijaiyah learning prototype that scores pronunciation on a scale.
Practice only improves when the learner gets feedback. In Quran reading, that feedback has traditionally come from a person who can hear the letter and correct it on the spot. The research team behind a Hijaiyah mobile learning prototype asked a practical question: how much of that loop can a phone carry on its own?
Before writing any scoring code, the team ran a field study with five participants who were already comfortable with Android apps and had tried Iqro level 1. They built personas and a user journey map to capture where the existing apps lost people.
Turning audio into a number is only half the job. The score also has to mean something close to what a teacher would say. In the feasibility test, ten learners were scored by both an Ustaz and the app. The two sets of scores moved together with a correlation of 0.51, and the app generally landed higher than the human judge.
The clearest benefit lands with learners who cannot meet a mentor often. A daily score, even an imperfect one, keeps practice moving between sessions. The authors also note that the method still needs more respondents, a fourth acoustic feature based on timbre, and trials at higher Iqro levels before it can be offered as an evaluation tool.
The wider lesson for learning apps is simple. Change the output from a verdict into a measurement, and the learner gains a reason to try again.
Komang Candra Brata, Mohamad Handy Nugraha, and Adam Hendra Brata (2022). An alternative m-learning scoring approach to automatically evaluate Iqro's pronunciation based on pitch, volume, and rhythm features. International Journal of Electrical and Computer Engineering.
