This post is also available in: Deutsch (German)
Studies show that a higher level of education is a strong predictor of longevity due to many factors. When Big Data Science can improve education, it will likely lead to a longer life. We discuss what is wrong with the current education system and how Big Data and Data Science can help in its digitization.
What is wrong with mainstream education?
“So how did my love for school change? Simple: school stopped being about learning,” student Isabella Bruyere wrote in her Medium blog post when she was a high school sophomore in 2017.
Regarding the flaws of school, or the education system as a whole, she couldn’t have said it better. This is the sentiment that many of us would share in common with her.
A listicle expressed how “school is making children stupid and depressed” in eight ways. One of the ways mentioned in the listicle that probably strikes a chord is point number 2: “It teaches children what to think, not how to think”.
“School does not teach children how to develop their capacity to think logically…
On the contrary, children are forced to believe in the things they are being taught…
Hence their critical thinking is prevented from improving, and as a result children are turned into stupid automatons.”
Freedom and Safety on point number 2 in the 8 Ways School is Making Children Stupid and Depressed.
This is such a major concern that TED talks about whether schools kill creativity have been delivered.
Image Source: Nikhita S
This heavily-discussed topic is accompanied by studies about the detrimental effects of assigning too much homework to children.
The National Education Association (NEA) and the National Parent-Teacher Association (NPTA) in the USA recommend assigning 10 minutes of homework per grade level and no homework for kindergarten kiddos. This means 10 minutes of homework for Grade 1, 20 minutes for Grade 2, 1 hour for Grade 6, etc.
However, a study found that students are forced to do three times the recommended workload.
“The data shows that homework over this level is not only not beneficial to children’s grades or GPA, but there’s really a plethora of evidence that it’s detrimental to their attitude about school, their grades, their self-confidence, their social skills, and their quality of life,”
A contributing editor of the study, Stephanie Donaldson-Pressman, told CNN.
But what does Big Data have to do with this? A lot. It means actually paying attention to students’ learning and psychological needs. As we’ve seen in Part 1 of our Big Data-driven sustainable development goal initiatives article series, Big Data has the capability to improve the education system.
Lots of data are collected in the educational system
First things first, e-learning publishing and knowledge-sharing platform, eLearning Industry, describes data-driven decision making in schools as “the systematic collection, analysis, examination, and interpretation of data to inform practice and policy in educational settings”.
In other words, educational data is collected and analysed to improve the provision of education in schools. Large amounts of educational data, especially in the higher education system, are gathered daily from various sources and various data origins in various formats. The various types of educational data according to eLearning Industry are:
- “Student data, such as demographics and prior academic performance.
- Teacher data, such as competences and professional experience.
- Data generated during the teaching, learning, and assessment processes, both within and beyond the physical classroom premises, such as lesson plans, methods of assessments, classroom management.
- Human Resources, Infrastructure, and Financial Plan, including educational and non-educational personnel, hardware/software, expenditure.
- Students’ Wellbeing, Social and Emotional Development, such as support, respect to diversity and special needs.”
Image Source: NeONBRAND
In a more specific example, Summit Learning, an online central equivalent of school filing cabinets, has a longer and more detailed list of data on its platform. Feel free to skip the list if you’re impatient.
- Application Technology Meta-Data: Automatically-collected data on its website such as visits, interactions, IP addresses of visitors, cookies and other browser settings that may help in identifying the source of technical issues and supporting user access and security.
- Application Use Statistics: Automatically-collected data about interactions with the website’s features to identify useful features.
- Assessment Data: They consist of test scores and observations to monitor the progress of students.
- Attendance: Data about daily school attendance, class attendance, suspensions and expulsions may be used to monitor truancy.
- Demographics: Each student’s birth date, gender, ethnicity, language proficiency, socioeconomic status, disability, special education needs, etc may be used to better understand the needs of a diverse student body and cater to them.
- Enrollment Information of Students: This includes a student’s grade level, homeroom engagement, mentor, enrolled subjects, and graduation year.
- Student’s Information: A student’s name, ID number and contact information.
- Parent/Guardian’s Information: A parent or guardian’s name, ID number and contact information, especially if the student is a little kid.
- Schedule: Student scheduled courses and teacher’s schedules have to be matched to avoid clashes.
- Student Performance: Online education performance, offline class work performance, tertiary eligibility, feedback from teachers, etc.
- Student interests: Extracurricular activities and non-academic achievements.
- Survey Responses: What the students, parents and teachers think about the provision of educational resources and approach to teaching and learning.
It seems that educational data aren’t limited to a student’s learning experience. The data also include a wide range of non-academic data for administrative purposes, which are also important for a student’s learning experience.
Imagine being told that you can’t sit in a certain class just because you didn’t register for it.
Image Source: Mikael Kristenson
Hence, all these data are opening doors for those involved in education to take advantage of Big Data.
What the education system can learn from Big Data explorations
The goal of the education system is to educate the students, so the abundance of data in the education system can inform several parties to help achieve that goal.
Who can access the data and how can different data help each party? The Data Quality Campaign (DQC), the voice for education data policy in the USA, sums this up perfectly:
- students: “I know my strengths and where I need to grow. I can shape my own education journey.”
- parents: “I know what actions to take to help my child on her path to success. I can be a better champion for her.”
- teachers: “I know where my students are succeeding and struggling right now. I can help them grow.”
- school leaders: “I know what’s working and what isn’t in my school. I can make timely decisions and make sure resources support great teaching and improve student learning.”
- afterschool partners: “I know what’s happening with these kids before 3:00 p.m. I can help families and communities create more opportunities for students to succeed.”
Let’s look at an obvious example of how Big Data eases the burden for everyone.
Real-time and automated progress tracking
In a paper about Big Data for Education by Darrell West of think tank Brookings Institution, a hypothetical scenario was used to demonstrate how a student’s monitoring of progress can be sped up by systematic, real-time educational data.
Instead of getting a teacher to manually grade test papers and hand them back to the students a few days, weeks or months later, students can instead go through a computerised learning software to study.
The students can then take a quiz that, upon completion, gives them instant feedback as well as recommendations to resources that will help them understand a topic in more detail for improvement.
The teacher would also receive an automated performance report from the software to know what needs to be improved.
Adding to this, the monitoring of progress can also be sped up by the automation of instant feedback generation using supervised learning algorithms, according to a paper by the Institute of Electrical and Electronics Engineers (IEEE).
Personalisation of learning for individual learning needs
At the same time, such mining, analyses and machine learning of assessment data can quickly determine the optimal approach to teaching and personalised curricula.
A paper by the US Department of Education about Enhancing Teaching and Learning through Educational Data Mining and Learning Analytics discussed the personalisation of learning for individual students using a hypothetical scenario.
In an introductory biology course with room for personalisation, the teacher’s role has changed from telling to designing, organising and supporting learning experiences. The teacher communicates the course’s learning objectives and suggests resources and experiences that can attain those objectives.
This is an alternative to the cookie-cutter approach of forcing all students to listen/doze off to the same lectures and do the same homework at the same time. So, the students go through the required resources and experiences using their own learning paths.
Image Source: Susan Yin
Focusing on a student who has reached a point where the next unit is population genetics, the student’s dashboard in an online learning system shows a variety of learning resources about population genetics.
The resources include lectures by a master teacher, awesome videos about genetics, interactive genetics simulation games (so fun!), a group project and revision material.
Each resource has a student rating for ease and enjoyment of use, derived from past activities of all students, including “like” buttons, assessment scores and correlations between student use and assessment scores.
The ratings are continuously updated as students interact with the resources. The updated ratings are then shown on the dashboard, indicating how much of the population genetics unit content, which the student has not yet mastered, is covered by each resource.
In other words, how much each resource can help the student master the unit and whether the resource has any room for improvement.
The teachers and schools have access to the online learning system data, making it easier for them to assess the students’ learning needs and certify the students’ accomplishments.
Relearning the way we learn
This year, the Covid-19 pandemic lockdown taught us a few important lessons and one of them is the need for online education, as the lockdown forced schools and universities to bring their lessons online.
A shift to online education might make educational analytics a lot easier to do as all the data can be extracted online. By learning about the behaviour of online education users, education institutions can figure out what to do next.
For this shift to happen with a higher likelihood of success, education parties need to focus on effectively using the data instead of just collecting the data.
I find that the Data Quality Campaign (DQC) is a great resource for this, so you can check out its proposal for data policy in more detail by clicking here. Summarising its proposal, DQC urges policy makers to:
- set questions that prioritise data analysis that can inform state action;
- make sure teachers can access and use data effectively;
- give timely, high-quality, relevant and accessible data to the public;
- communicate the types of data collected, their value in supporting student learning and how data is protected;
- set privacy protection and confidentiality policies.
Perhaps, this approach can also applied to educational institutions in the rest of the world. The large data volume collected in education systems worldwide makes the education sector an attractive opportunity for Big Data analysis.
I hope that this article has proven that Big Data Science is capable of driving decisions regarding administration, identification of and adjustments to various learning needs’, organisation of resource allocation, teaching performance and many more measures to take to support student learning.