Speaking was one of the most essential skills in English learning, yet it remained
difficult for many students, especially in EFL contexts with limited exposure and practice.
In Indonesia, speaking classes were often teacher-centered, which led to challenges in
fluency, vocabulary use, and confidence. At SMKN 4 Palembang, these problems were
also evident, as students showed limited speaking performance despite frequent use of
digital media such as YouTube in their daily lives. This study investigated the correlation
between self-directed learning on YouTube and students‘ English speaking achievement.
A correlational quantitative design was employed, involving 33 tenth-grade students
selected purposively. The instruments consisted of a questionnaire that measured the
intensity of self-directed learning through YouTube and a speaking test that assessed
fluency, pronunciation, grammar, vocabulary, and comprehension using a rubric adapted
from Brown (2004). The results showed a correlation coefficient of r = 0.116, indicating a
positive but very weak relationship. The significance value was p > 0.05, suggesting that
the correlation was not statistically significant. These findings revealed that while
YouTube was widely accessed by students, its use as a self-directed learning tool did not
strongly contribute to speaking achievement without structured strategies and reflective
learning. Future studies were recommended to integrate YouTube with guided learning
approaches, extend the duration of practice, and provide clear learning objectives to
maximize its potential for improving speaking skills.
Keywords: Self-directed learning, YouTube, Speaking skills, English, correlation