By Bronte H. Lacsamana 

The Filipino students who scored poorly in reading comprehension in the Programme for International Student Assessment (PISA) in 2018 mostly come from low-income backgrounds where family and school environments do not motivate growth mindsets, according to the findings of a machine learning (ML) study by the Dr. Andrew L. Tan Data Science Institute (ALTDSI) of De La Salle University (DLSU). 

Eighty percent of the Filipino students, all aged 15 years old, scored below level 2, PISA’s recommended minimum proficiency level. Of these, 83% came from public schools.  

“Students belonging to lower income groups may find academic requirements as a burden as they prioritize work to help their families earn a living,” said Rochelle Irene G. Lucas, Chair of DLSU’s Department of English and Applied Linguistics, at a webinar where the results were unveiled. “To them, reading becomes a task solely associated with school requirements and considered as just a waste of time and resources.” 

Though the relation between socioeconomic background and educational attainment is not new, she added that the data science findings confirmed the important role of a child’s environment and available resources in motivating them to learn and improve.   

“Reading is substantial in a child’s understanding of concepts in math, science, and other subject areas, so that’s why they tend to score low in those areas as well,” she answered, on why the study focused on reading comprehension. 

Macario O. Cordel II, dean of ALTDSI explained the need for a data mining algorithm based on computational game theory: “If you look at the data set, you have thousands of variables to consider, and then thousands of student data that you need to crunch.”  

He and his team also revealed a glimpse of data visualizations grouping the students by region on the Philippine map and providing relevant information about their learning environments and scores. They plan to publish the visual map online in the near future.   

The study was able to cluster 20 variables in a student’s environment into four factors: reading, teaching, ICT (information communication technology), and motivation. 

The reading variables include negative reading self-concept, low awareness of reading strategies, low enjoyment of reading, and low reading of fiction for leisure.   

“These are students who have already identified themselves as having difficulty in reading, which ties into their motivation,” said Allan Benedict I. Bernardo, Chair of DLSU’s Counseling and Educational Psychology Department.   

He added that teaching variables — frequent teacher feedback, asking students their thoughts on the reading material, and low teacher enthusiasm — did little to help. The negative feedback and tendency to put students on the spot often demotivated them.  

Meanwhile, the lack of ICT resources, common in public schools and low-income households, provided no opportunity to make learning interactive.  

“Access to ICT can be an important entry point in changing experiences of the family,” he explained. The variables under ICT were low ICT resources at home and infrequent use of ICT to learn about topics, chat with classmates, or read e-mails.  

Finally, motivational variables proved the need for socio-psychological focus on students as well, according to Mr. Bernardo: “What our research shows is that if we focus on the social and psychological experiences of these learners, we predict in a very high level of accuracy who are the poor readers and who are the better readers.” 

The motivational variables include low persistence in mastering tasks, low mastery learning goals, low valuing for schooling, low expected occupational status after high school, and low growth mindset beliefs. 

English and linguistics professor Ms. Lucas suggested that, with the family as a child’s major influence, parents could be taught to nurture reading habits in the household. 

“There was a reading caravan that we did a decade ago which also taught parents how to read to their kids,” she shared. “They were able to read together in their free time.”  

As for improving the current curriculum, Mr. Bernardo noted the efforts of the Department of Education: “DepEd is now very much focused on the curriculum as a response to PISA and there has been a long-standing interest in focusing on instruction, but we should not be content with a ‘one size fits all’ type of intervention.”  

Aside from reaching students and their families and tweaking the curriculum, the research team implored policymakers and other education stakeholders to take a look at the bigger picture.  

“Even if motivation resides in an individual, the socioeconomic model suggests that these motivational sets do not occur in a vacuum. They are adapted within these communities,” explained Mr. Bernardo.  

He added that, in the larger social environment, if celebrities and politicians became models of success in spite of low academic achievement, it would foster a culture where education is not important.  

“The students are not blank slates. They’re making these motivational choices based on what they see — in school, in their families, on TV, in larger society, in their barangays. It’s not just working on the student, but on the environment itself,” he said.