Abstract:

This research investigates the role of demographic factors in adult learner performance in data literacy courses, a relatively unexplored area in educational research. The study analyses the records of 680 learners from the Data Literacy Program Basic (DLPB) across eight cohorts at the National University of Singapore (NUS). Firstly, this study aims to explore the demographic influences on learner performance and aims to build a predictive model that can help to identify learners at risk of falling behind and to provide them with additional support. Secondly, our study aims to enhance the learning effectiveness among adult learners by investigating class group demographics to help ensure demographic diversity across the class groups so that all classes will have adult learners with similar learning capabilities. The uniform demographic diversity will aid instructors with their time management in the classroom. Thirdly, our study aims to enhance the diversity of the demographic characteristics in the project groups to ensure project groups are balanced in terms of the demographic variables and to prevent some project groups of having a considerable advantage to other groups because of their demographic make-up.

The research examines the impact of ten demographic variables and other course information on learner performance in the various assessment components. Utilising Python programming a Multiple Linear Regression Model (MLRM) was built to predict learner performance at the end of the course based on the learner’s profile and assessments at midterm. Pearson's Correlation Coefficients (r) and the Chi-Squared Test was performed to identify significant associations between the adult leaner’s performance and the demographic characteristics. Further, a Mann-Whitney U-Test was performed to test whether there was a significant difference in the Project Group marks based on the Unit and the Final Grading based on the Unit.

The Pearson Correlation Coefficient showed that the E-Learning Quiz (ELQ) (r=0.73) and Take-Home-Assignment (THA) (r=0.86) were highly associated with the Final Grade (FG). Further, the Chi-Squared test results showed that highest education (p=5.318e-05), number of children (p=0.0286), project groups (2.786e-06), has significant association with the adult learner’s performance. In addition, the Mann-Whitney U-Test showed that the performance in the project assessment and the final grade was significantly different for different Units. The MLRM results indicated that ethnicity, highest education, STEM, Group Size, ELQ, In-Class Quiz (ICQ), THA, and Group Project (GP) were significant factors for predicting the final grade. The Mean Absolute Error (MAE) for the MLRM was excellent, 2.75.

This study emphasizes the importance of incorporating demographic diversity in educational programs to improve learning outcomes for adult learners. It suggests that recognizing and supporting learners at risk of falling behind, organizing data literacy classes and project groups based on demographic characteristics, can lead to more balanced and effective learning and teaching environments. The findings of this study are valuable and actionable as demographic imbalances can be addressed to enhance allocations of learners to classes and project groups and thereby create a supportive learning atmosphere for successful adult learner performance in data literacy.

Keywords:

Adult learners, data literacy, demographics, learner performance, educational achievement, prediction model, effective learning, effective teaching.