Use of Augmented Analytics in Educationby Becton Loveless
For as long as teachers have existed, they have performed data analysis. The role of a teacher demands it. The earliest teachers asked questions of their students to see their lessons were being communicated correctly. Tests have been used for ages to better understand to what degree students have mastered their lessons. In the 20th century, standardized tests have been analyzed to determine where students rank among their peers across the country. However, more than ever, educators and school administrators need to understand how data analytics can help them in their week to week student assessments.
An overview of data analytics
To begin with, it’s important to understand what data analytics is, precisely. Data analytics is a science designed to use raw data to create insights in to a situation or phenomenon. In other words, it’s easier to understand a situation when a scientific review of the existing data has been conducted. Data analytics is used to see trends in raw data and is often performed using a computer algorithm. Once trends have been identified in the data, organizations are better equipped to devise strategies that improve their performance.
Just about any type of information can go through the data analytics process. For years, data analysis has been used by businesses to help them see trends in their performance and develop ways of improving that performance. Businesses, for instance, can see how certain machines are performing or how much downtime workers are experiencing due to situations in the workplace. This information is revealed when data analytics uncover trends in the data showing slowdown at certain points in the workflow. Data analytics can also be used to see how certain products are performing and keeping consumers engaged with those products.
There are four types of data analytics used by various organizations.
- Descriptive analytics describes an organization’s performance over a period of time. This form of analytics shows whether an organization is meeting its objectives.
- Diagnostic analytics takes data and attempts to find out what happened. It’s possible that unique circumstances may impact performance during a certain time of the year, for instance.
- Predictive analytics try to predict what will happen next. On the basis of previous performance, this form of an analytics is an attempt to reasonably predict how trends will unfold in the future.
- Prescriptive analytics takes descriptions of current conditions, diagnosis of why things happened the way they did, and predictions of the future to prescribe a way to address current issues.
While it is true that analytics have been used in businesses and various industries for a long time, education researchers have slowly started to point out the many benefits of applying analytics to the field of education.
Kinds of Data
In schools, there are two forms of data used for two distinct purposes. The first form of data is administrative data, which includes behavioral, achievement, and demographic data. This data is collected not only from schools, but also from government agencies and other organizations involved with collecting such data.
Administrative data is collected over long periods and includes some of the largest data sets possible. Since the data involves so many people and is collected over such long periods, it’s often presented in a yearly report. Examples of the data included in this report range from census data to standardized test scores.
The second form of big data includes learning process data. This kind of data also includes many participants. For schools, this means data is drawn from many students. However, there’s also a good deal of data that’s collected about the individual students as well. This data is generated from many different observations of how the student is performed and coded so that trends can be more easily identified in the data. Learning process data can be collected using commonly available technology, such as online assessments and interactive technologies that are increasingly more popular at both the primary and secondary level.
These separate data sets allow researchers and educators to identify different types of trends. Administrative data is well suited for presenting how schools, districts, and even the entire education system is performing. Not surprisingly, administrative data is more useful for administrators from the individual school level up through government officials involved with the oversight of the education system. Learning process data is far more useful for teachers, since it can provide some insights into how a classroom is performing.
The Hypothetical Classroom
In a report for Education Week, Benjamin Hold detailed how big data and analytics will eventually be applied to education and student learning. In his article, Hold portrayed a classroom in which all aspects of the student’s day is monitored. This extreme version of a classroom would include infrared cameras documenting everything a student touches, cameras recording facial features and fidgeting, and wearable devices that track student heart rates and other physiological signals.
This isn’t likely to be the future. Most teachers and students will immediately see a number of ways that such a classroom might be problematic. From the invasion of privacy to the constant feeling of being monitored, such a classroom might bring with it a whole host of negative qualities that would hamper the learning process. Of course, parents are just as likely to be upset with what comes across as a surveillance environment created in the classroom. For all these reasons, it’s not likely that this version of data analytics will ever be applied on a large scale in the American education system.
However, there’s a lesson to be taken away from this extreme version of the future. At some point, big data is going to become very important to the learning process. Big data refers to large data sets that can help identify trends and relationships. Considering that data analytics performs better when there are more sources of data, big data and data analytics go hand in hand. In the hypothetical classroom discussed by Hold, big data sets are generated not only from student performance but from their individual habits, physiology, various behaviors, and more.
In reality, the data generated for educators will most likely be much less invasive moving forward. However, it will still require the aid of computers to help make sense of the data being generated.
Current Big Data and Data Analytics
For some time now, public schools have been generating data at levels that would have been impossible in the past. Learning software has made it possible to track a student’s performance and determine bottlenecks where a student has begun to struggle with their work. This kind of data allows teachers to tailor new teaching interventions to students once they find out where the student is having difficulties.
On a much larger level, school districts are able to gauge their overall performance by compiling all this individual data into large data sets. School districts can determine whole school performance and attendance rates. In some cases, school districts exchange these large data sets with other public agencies to better understand how they are performing when compared against other districts. This exchange of data makes it possible for schools to make predictions about which students are at the highest risk of dropping out or simply becoming disinterested in their work. After determining who is at risk, interventions can be introduced that address the needs of these individual students who are struggling in school.
Data Analytics for Teachers
In many cases, when technology is introduced to a school, educators often settle for using it at the most basic level. Technology is often used to meet basic teaching requirements and teach the core essentials of a subject. However, there’s the potential to use technology in much more sophisticated ways. Data analytics can help teachers better address the needs of their students by finding where students are weakest and addressing those shortcomings.
Big data can help teachers to adjust their teaching in real time to address a student’s struggles. By applying analytics to data sets generated by a student, teachers can identify whether a student is struggling with their material. If they are, then teachers can adapt their teaching styles to improve student performance. Teachers often bring unconscious biases to the classroom and favor certain teaching methods, but data analytics can help teachers understand if these methods are being as effective as they think they are.
Data analytics can also help teachers settle on what materials they give to students. For instance, when teachers move on to a new section of material, students find themselves learning new lessons that some of them are better equipped to understand. Some students struggle in the previous lesson, which compounds their difficulties in learning new material. Teachers can identify these struggling students and give them different starting materials when a new section of the course begins. For struggling students, these materials can focus more heavily on the previous lesson and more clearly outline how prior material is connected to new material. Students who were proficient in the previous lesson can begin the new section of the class with more advanced materials.
Computer Based Assessments and Analytics
The benefits of using data analytics is far ranging, but how is data analytics actually performed by schools? One of the most common methods is by using computer assessment. Historically, traditional assessment has involved gauging student knowledge using tests, quizzes, and homework. Teachers would then review these assessments and identify students who were falling behind.
However, machine assessment has become increasingly common in recent decades. Specifically, computer adaptive testing (CAT) has become very common in schools. Researchers Bill Cope and Mary Kalantzis described just how effective CAT could be at improving the performance not only of students, but of teachers as well. The CAT approach represents course-level data that can be acted upon.
For teachers, course-specific data is the most valuable. Larger datasets tracking the entire school benefits administrators, but teachers need to receive timely data about the courses they teach. Data is useless if it can’t be converted to usable knowledge in a relatively quick timeframe. For that reason, student outcomes on assessments and activities need to be interpreted quickly. This data needs to be concrete enough that teachers can plan a course of action by which they can address shortcomings in the classroom.
CAT testing adapts to students as they demonstrate a certain level of knowledge in a topic. As the student progresses through their material, the computer responds by delivering harder and harder questions if the student is performing up to certain expectations. However, the computer can also deliver easier questions if the student is struggling. Because the test is adaptive, the potential for cheating declines significantly. These tests also help teachers to more accurately understand a student’s comprehension of a specific topic.
The data analytics portion of CAT comes into play once a review of sub-sections of an assessment begins. Different parts of these computerized tests are coded. So, while the tests produce an overall score, they also produce sub-scores on different parts of the exam. Teachers can not only identify sub-sections where a single student is struggling, but sub-sections where an entire class is having difficulties. It’s one thing to say that a class struggled on a test about marine biology. However, it’s far more useful to be able to identify what parts of marine biology the class struggle in. Students may find it hard to learn about whale biology, for example.
Coded parts of tests help identify trends in the data and reveal where the entire class is struggling. This not only helps a teacher identify what material they need to review but also helps them identify parts of their material they might want to teach differently in the future. Coding of test sections doesn’t require computer testing, of course. It’s simply that computers allow for coding and data analytics to be performed with much more effectiveness.
Computer based assessment does more than measure content knowledge however, and for this reason they have an advantage over traditional analysis. When students take their courses over a computer, the software can also track how long it takes for that student to go through the material and master it. This provides teachers with valuable information about where students are struggling. Using computer-based assessments, teachers can not only see what a student got wrong, but how long it took them to get things right. The ability to analyze the length of time required to achieve mastery is important to teachers because it allows them, once again, to see whether the entire class is struggling or if the problem is restricted to only a few individual students.
The application of data analytics to the field of education is still in its infancy. While businesses have long used data analytics to improve their performance, there are still questions about how to use such an approach when teaching students. Questions about privacy and invasiveness abound, but there’s a balance to be found between too much data and too little. With the right amount of data collected from various sources, data analytics has the potential to improve not only the academic performance of students but the teaching methods of instructors as well.