Article Sidebar
Main Article Content
Abstract
The experimental study involves a qualitative methodology, which has been strengthened with machine learning, to investigate the success predictors of mathematics among the secondary school students in the Nigerian region of Ebonyi that have been characterized with socioeconomic issues and inequality in education provision. Semi-structured interviews and focus group discussions were used to collect primary data where 150 participants (80 students, 40 teachers, and 30 parents) were engaged in semi-structured interviews and focus groups with a purposive sample of participants to provide a representative sample of these groups in relation to their demographic factors (gender, age, socioeconomic status, and rural-urban delimits). The transcribed information was subjected to machine learning, with the latent Dirichlet allocation model utilized to model the topic analysis and Vader used to analyze emotions. Its demographic composition presents 51 and 49-percent men and women with an average age of 16 years old, 53-percent living in low-income households, and 55-percent in rural areas which correlate with the overall population data in Ebonyi State where the adult literacy rate is at 68.1 and the completion rates of primary education are at a lower level than the national rates. It has shown that the quality of teachers, motivation of students, the presence of resources, and the support of the family are leading factors, and positive feelings were related to intrinsic motivation (average 0.28) and negative to the infrastructural shortage (-0.38). Basing on the activity theory, the paper views these factors in terms of a system of learning activity with contradictions between the agency of individuals and unfair systemic regulations. This mixed method approach not only is more scalable to the qualitative analysis, but it also offers the perspective on how to tackle the problem of under achievements in mathematics, which is paramount in sub-Saharan Africa where the percentage of passing grade tests in the national examinations are around 40 per cent. Policies such as teacher training based on the use of digital tools and community intervention programs can be recommended to support the motivational and resource structures. The study will make contributions to the body of educational research because it proves the usefulness of artificial intelligence in streamlining qualitative investigations to achieve fair STEM results in underdeveloped settings.