A Brief Look at Analytics Today
The reason why Netflix is so valuable as a company isn’t just because of its large subscriber base, it’s because the company knows exactly what kind of entertainment media you like to consume in detail down to the words that the actors say. According to this article, Netflix tracks data points like exactly when movies become too scary for you. No wonder why so many hit series and shows are now produced by Netflix, they know what you like better than you do.
If you think this is unsettling, think also about why social media can be so addictive, companies like Google, Facebook, and Tiktok track exactly what makes people stay longer on their platform, and personalizes content for people to maximize the time they spend on the platform. After all, they make money by selling your attention.
Big data and analytics have been around for a long time. Companies like SAP, Accenture, Tableau, and Oracle have been in the business of analytics for over a decade. And while Cambridge Analytica ceased operations in 2018 after its scandal with Facebook, its parent company SCL Group, founded in 1990, is still in operation.
Analytics and business intelligence are, then, not a new field by any means. However, with the rise of alternative education platforms such as MOOCs, the introduction of learning management systems in the classroom, as well as software applications complementary to the learning experience, this digitization of learning is accompanied by a large influx of big data. This data can only be analyzed computationally because of its sheer size. All of this has led to a rise in the field of analytics for education. This can be seen in the form of companies that analyze learning data to allow decision-makers to make data-driven decisions or software learning solutions that make use of learning analytics to directly intervene in the learning experience.
Learning Analytics Explained
In terms of analytics for education, it’s important to note that we will focus on learning analytics and not academic analytics, which is a form of analytics that concerns itself with admissions, retention, and other educational data points that do not directly have anything to do with optimizing teaching and learning. The definition of learning analytics is as follows:
LEARNING ANALYTICS is the measurement, collection, analysis and reporting of data about learners and their contexts, for purposes of understanding and optimising learning and the environments in which it occurs.
As defined by the Society for Learning Analytics Research (SOLAR)
This short video by SOLAR will give you a good foundational understanding of learning analytics, a type of learning analytics that concerns itself with improving teaching and learning, pedagogy, and behavior inside and outside of the classroom.
There are various types of learning analytics, each with its own purpose:
- Descriptive analytics looks into the data of the past and seeks to explain what happened. An example of this would be adjusting teaching practices based on course feedback and evaluations.
- Diagnostic analytics takes it a step further and uses data mining and correlations to figure out why something happened. An example of this would be an in-depth analysis of whether having an assignment due on a certain day of the week affects student scores on the assignment.
- Predictive analytics identifies patterns in past data in order to capture relationships and predict trends for the future. An example of this would be using regression analysis to predict the likelihood of the student becoming an “at-risk” student based on data inputs of demographics, past behavior issues, and academic records.
- Prescriptive analytics uses both descriptive and predictive analytics to recommend a course of action toward an identified future outcome. An example of this would be a program with the ability to recommend specific practices and interventions that can improve a student’s reading comprehension based on data points of their past performance and progress.
When understanding learning analytics, it’s important to note that applying analytics to your own classroom does not have to be complicated. All that needs to happen is the gathering, visualization, and analysis of data that will help you as an education professional. This could be as simple as recording your students’ responses on a multiple choice test, learning which questions the students struggled with the most and using this information to plan lessons that reteach material students struggled with. Applications like Zipgrade have built-in functions that allow teachers to do this. The whole point is about using data to support your decisions.
Ventures Making Use of Learning Analytics
The learning analytics landscape is set to boom. With the global pandemic as a catalyst, online degrees, MOOCs, and other educational marketplaces have become very popular. Many traditional classrooms have adopted digital classrooms, and technology has become an integral part of the learning process. Data, lots of it, is a by-product of this occurrence. The ability to visualize, analyze and help educators make meaningful decisions with this data is at the core of what will make a learning analytics venture valuable. To help you better understand how learning analytics have been used in ventures so far, I’ve listed two companies that make use of learning analytics in their products. What’s interesting to note is that both of these companies are decades old. The field is emerging but not new. Take a look at the videos and think about how you can apply learning analytics in your own context.
Renaissance Learning has an Accelerated Reader that uses analytics to match students to personalized reading material.
Dreambox Learning uses analytics to provide personalized mathematics and English practice to students in real-time. Learners treat the process as a fun game, and learning becomes self-regulated as a result.
Inhibitors to Industry Growth
The main bottleneck to learning analytics integration into schools and classroom lies around the collection, analysis, and interpretation of data. Most ventures are like Dreambox and have built-in processes for data collection and analysis. However, this means that the data is proprietary to the venture and not available for schools and educators to analyze. If we want learning analytics to be integrated into teaching and learning, we need the data gathered to be holistic and representative of the learner’s profile and a place to centralize and analyze learning data. A few challenges prevent this from happening:
- Teachers lack training on practices concerning data. Educators would need to be at the forefront of data collection and recording. While there are applications that aid in data collection, many teachers lack the technical expertise to deal with data. Much education needs to happen in educating teachers about the impactfulness of data-driven decision-making, and how they can get comfortable collecting and analyzing data in their practice.
- There’s a lack of standardization among collected data. In order for data to be meaningful, we need to be able to look at datasets across learners and classrooms and be able to make decisions. Because each classroom is inherently different, and each teacher’s judgment is inherently different, it is difficult to collect organic data that can be applied to different contexts without accounting for the effects of biases during the data collection process. Learning data collection and treatment frameworks need to be established, standardized, and implemented across educational institutions.
- There are ethical and privacy issues around data collection. The sheer amount of data that would need to be gathered is massive. When dealing with K-12 learners, data collectors need to be clear on what data should be gathered, what shouldn’t be gathered, who should have access to the collected data and how it is stored, how informed stakeholders should be, and whether informing stakeholders will compromise the quality of data being gathered. Guidelines on ethical data use, collection, and storage need to be taught to relevant stakeholders.
- Data is currently being collected and stored in a fragmented way. It’s difficult to look at the big holistic picture and apply learning analytics in a holistic way. Different teachers and schools use different educational technology products, learning management systems, and student information systems. There’s no central storage for learner data, making it difficult for any individual party to paint a representative picture of the learner using data. Firms, governments, and individuals need to cooperate to make learning analytics more capable of optimizing teaching and learning.
While these challenges currently act as inhibitors to industry growth, they also represent problems for budding entrepreneurs and intrapreneurs to tackle. What’s true is that the amount of learning data being collected grows by the second, and so the venture that is able to unlock the mysteries of this data to actually optimize teaching and learning will surely experience major tailwinds.
References
Charles Lang, George Siemens, Alyssa Friend Wise, Dragan Gašević, Agathe
Merceron (Eds.). 2022. Handbook of Learning Analytics (2nd. ed.). SoLAR, Vancouver,
BC. DOI: 10.18608/hla22
About Me
I am an economics and business educator currently teaching internationally. My interest in ed tech came from having the opportunity to work as a project manager for a proprietary student management and learning management system in my previous school. During my time there, I was a champion for using demographic, behavioral, and assessment data collected by the system to build predictive analytics for identifying potential ‘at-risk’ students.
Discussion Prompts
Now that you have a basic understanding of learning analytics and its applications. Respond to one or more of the following prompts below as a reply to this post:
- As learning analytics software becomes more prevalent in the educational landscape, how do you think mass adoption of learning analytics software above will change teaching and learning? What are the risks and benefits of school districts adopting software like Dreambox and Renaissance?
- Many educators actually already apply the concepts of learning analytics in their own contexts. After all, educators are constantly making teaching-related decisions based on the feedback they receive from students in the classroom, and on assessments. How are you already utilizing the principles of learning analytics? How can you take a more formalized approach and apply learning analytics to inform your professional practice in your own context?
When you’re finished, you have a couple of options. Go back to the launchpad post. Go read the post on big data.
Hello Leon, thanks for all the great information introducing us to Learning Analytics. I appreciate the many examples you included in the post and summary of some of the limitations in the industry. In response to your discussion prompts: 1) Regarding the mass adoption of learning analytics in teaching and learning environments, my main concern is some districts having the financial and personnel capacity to utilize these types of tools, and some not. As you mentioned, infrastructure and learning for educators needs to be put in place, and in areas without capacity, I feel like their educators and learners may be left in the dust. This will not work towards “levelling the playing field” for learners, which is so important to reduce the participation gap in education. I think there are many benefits to adopting innovative software like Dreambox and Renaissance, but it’s hard to imagine these tools being adopted largescale and reaching those who may benefit from it the most. 2) I am glad you mentioned that learning analytics does not have to be complicated, and that many educators are doing their own participatory research and data collection that helps inform their teaching. I do not work in a traditional classroom setting, but in my current position we use data analytics to look at how many people are using our educational resources, and we have used short pre- and post training surveys to get more information about how much was learned, perception of effectiveness and utility, etc. You definitely have given me some inspiration to think of how I can take a more formalized approach to this – so thank you for that!
You’re totally right. The financial costs of new technology will always widen the opportunity gap. In the same way that only some students will have access to certain extra-curriculars, or out-of-school tutoring, only some students will have access to new technologies or be supported by it. Even when the intention is to help everyone, localized integration is still likely to occur in the beginning. Decision makers need to see the value of data and learning analytics before even considering software and funding. The problem is they probably never will because of a lack of software and funding. I think the mindset of an analyst is super appropriate for your role and there’s a lot to learn. Analytics isn’t new by all means. It’s just relatively new when applied to education. I’m sure the first institution or individual to crack the code of the variables to educational success or failure in different contexts will reap awesome rewards.
Hi Leon, thanks for your thoughts, and I think as analytical techniques in education increase in popularity there will for sure be more efficiencies and innovations that are more accessible to wider audiences that may not have the capacity to be at the forefront of these technologies.
Hi Leon,
Thanks for all of the learning about Learning Analytics – it was especially valuable for me to consider the different types of analytics and the inhibitors you discussed!
I have used all four types of Learning Analytics, but mostly Descriptive and Predictive as a classroom teacher. I had my students give feedback on teaching, assignments, and classroom culture, and that subjective and descriptive feedback helped me to make adjustments as needed. I also used lots of predictive analysis, but less formally than as it’s described above. For example, I would look at documentation and talk to previous teachers to determine which specific students in my class might need certain supports and interventions, but would take a more holistic view rather than conduct formal regression analysis. In my current role as an assistive technology educator I think I use more Prescriptive Analytics, as I look at both descriptive and predictive analytics to determine which assistive technology and which specific practices and interventions for implementation will be best suited in supporting students.
I see the need to take a more formalized and data driven approach, but the inhibitors that you mention in this post are real and difficult to work through. Specifically #1 – not only do teachers lack training in data collection, they also lack time to train, collect, and analyze such data. Currently different teachers collect and analyze data in different ways according to capacity and district/school/parent/personal opinions of what is best fit. Renaissance, Dreambox, and other software/programs can be an effective support as you mentioned, but I agree that there needs to be understanding and involvement at the school level in terms of what data is being gathered and how it is being applied/used. I think there needs to be systemic change (as in most areas of education) for teachers to have time and capacity to address all of the aspects of learning analytics that are expected of them.
The education sector needing systemic change has always been a problem that’s been underserved. I agree with you in that a problem I see with analytics is whether we can actually treat unique students like variables and numbers. That’s a mistake coming from overreliance on the data. A regression analysis could complement or reinforce an existing hypothesis. It would be interesting to see if you can identify a set of indicators that can predict the likelihood that a student may need supports in the future. That kind of data would be useful to boards and districts and might actually be a springboard for funding teacher education in data and learning analytics. As for the systemic problem, I don’t think that can be addressed until the current generation of teachers who are comfortable with data and analytics become administrators in the future. Thank you for your sharing!
Hi Leon, thank you for sharing the details of Leaning Analytics. I’ve learned a lot after reading through your post, specifically the different types of learning analytics. It’s interesting to know that educators such as myself are doing learning analytics in our day to day teaching without sometimes even realizing it.
Discussion Prompt 1: I think personalized teaching, which is provided with learning analytics software is definitely the most efficient way to run a classroom. It could raise engagement and motivation for both students who are struggling and for students who need a challenge. Like the Dreambox video mentioned, one teacher cannot support all the students in their class and their varying levels of learning. Learning software with learning analytics built in can provide that. In an ideal world, this would be the case. However, with the lack of funding and resources in public schools it’s unlikely applications such as Dreambox or Renaissance will be adapted on a district level. By adapting these applications, the role of the teacher will change as well. Instead of a “lecturer” or an “instructor”, their job will become a “troubleshooter” or an “application manager”.
Discussion Prompt 2: I run a multi-year academy at the high school I work at. Every school year, I conduct a couple surveys at the beginning of the year and at the end of the year. One of the survey is a review of the year. It gathers information on engagement with the course contents, the assessments, and myself. This will gather data to help me decide on what course contents to keep and what needs to be revamped. It also allows me to adapt my teaching style to better engage with my students. The other survey I give out is at the beginning of the year. It gathers information that will inform me what my student population’s interests are, both inside the classroom and outside of the classroom. This information will help me create a curriculum that engages the students and allow them to engage with the contents. On top of that, knowing the students’ interests outside of the classroom will help me connect with my students and build a good professional relationship.
Hi Leon,
I’ve learned a lot about learning analytics through your OER. Compared to the other emerging technologies we have seen, I would consider it to be one that is less obvious in the sense that it is built into programs that are performing the analytics, it is challenging to understand the inner workings of a machine. As you mentioned, it is important for educators to understand more about the data collection, analysis and interpretation process to better guide the teaching and learning process. I think one of the greatest barriers is the standardization across data collection methods where it can range from the conventional grade books or anecdotal notes on sticky notes to online student portfolios like Freshgrade (which shut down in August 2022) after a large roll out/ push from the school boards. The power of technology lies in more than the collection piece and in the analytics of what to do with that information for students. It becomes the teachers responsibility to respond to the data to differentiate the instruction and find resources for the students to succeed- this is extremely time consuming and technology has proven to circumvent the process. In terms of how I am utilizing the principles of learning analytics in my own practice, I would argue that I am on the more conventional side as a Kindergarten teacher. The documentation and assessment process differs from other grades where it relies more heavily on anecdotal observations and holistic growth rather than on grades, scores or percentages. How I store and collect that data has changed because of technology over the years from an organizational standpoint (no longer is my binder overflowing with notes). Personally, I would argue that I use more prescriptive analytics, not necessarily with technology, to determine the best course of action for individual students. Again, it is difficult to implement a consistent piece of technology even at the school level, one that all teachers have access to. In the past we have used large bristol boards with levels to track student reading progress and to identify struggling readers. I’m sure there is a more efficient method of collecting and sharing this data with students that our schools have not yet embraced.
Hi Leon,
Thanks for your in-depth OER on Learning Analytics! Looking at discussion prompt one, I think there are a few barriers in using learning analytics software such as the accelerated reader tool and dreambox. I think the idea of an increased focus on learning analytics needs to come from the education level for educators working to complete their respective professional programs. Learning Analytics can be seen as a complicated topic even though many of us go through the thought processes without even realizing it. I think educators need to be given an opportunity to learn of the importance of it, but the largest issue is there isn’t enough professional development time. If aspects of learning analytics and using them in the classroom were taught to educators during their professional programs, there is a higher chance of it being used and becoming more prevalent in the education landscape. This is an incredibly important topic as it will help create more personalized learning which meshes will with the current BC Curriculum. I believe even if learning analytic tools were introduced at a district level, many educators would struggle and have trouble modifying a lot of their teaching methods. For that reason, I really think it would need to be introduced early in an educators career for it to have the largest impact.
Hi Leon and Alexei!!! Great post and great information! Thank you for sharing Dreambox and Renaissance they both look like great tools!! I agree that one of the inhibitors is “Teachers lack training on practices concerning data.” Therefore the risks of mass adoption may be that the data is used incorrectly to make the wrong insights or decisions. If the data is misused or misinterpretted, learners could be at risk of not getting the help and support they need to succeed. The benefits of Dreambox and Renaissance appears to be the personalization of the learning. Having learners grow along their own path at their own pace and giving an individualized learning experience. We are utilizing the principles of learning analytics through our new PowerBi dashboard and Pendo (a usage and utilization analytics tool). Through these tools, we are able to present and make decisions about our content and our software as a whole, we can understand what are the most complicated tasks for users and provide them with the learning and training they need at the right point within their process.
Thank you so much for your engaging post! In response to prompt #1, one thing that really excites me about the mass adoption of learning analytics is the individualized learning aspect that you have so clearly identified here. Learning analytics makes it possible for learners to be guided in an iterative way through learning materials that are relevant to them. One thing that concerns me a little about this mass adoption is what happens in other cases of data-driven algorithms. On social media for instance, algorithms have clearly siloed individuals into content pools and create kinds of feedback loops. Could these concerns arise as well in teaching tools like the ones your’ve mentioned? I would be interested to hear what you think!
I experienced a school system that was driven by learning analytics when I worked in the UK. Every student’s level of progress was predicted for the year using learning analytics and their expected level of progress based on past performance was shared with the child. At 11 years old students knew if they were not expected to make as much progress as some of their friends and were split into different sets based on their current levels and anticipated level of progress. A student in top set may feel quite good about this, but I felt like immeasurable damage was done to a student’s view of themselves as a learner when they realized that they were underachieving and that so little was expected of them. It impacted class dynamics in a way that made students in lower sets less likely to succeed as everyone in their class knew that so little was expected of them. They were of course told to keep working hard to exceed their predicted level of achievement, but that information seemed to become a self-fulfilling prophecy in many cases. The rewards of adopting software like Dreambox and Renaissance with the ability to gather data on student progress can only be realized if proper supports are available to students who are struggling. Adaptive software that is driven by learning analytics can be an invaluable tool in the classroom and can be beneficial for an inclusion model of learning as every student can access content at their level while the teacher works with smaller groups. I am very much in favour of learning analytics being used to power adaptive programs and provide information to classroom teachers but am less hopeful about it being used to drive data collection for school districts.
Hi Leon,
Thank you for your great work on learning analytics. I took learning analytics in the MET program last year and this is a great a refresher for me. In response to your first question, learning analytics can potentially helps school to deliver a richer learning experience for all students through differentiation. That way, public schools would be able to teach students to their strengths and better prepare them for successful careers later in life. And teachers would be able to inform parents, based strong data, why their children are on a personalized educational path.
The biggest issue I can think of would be to get enough funds in order to implement this kind of infrastructure to assist both teacher and students.
Thank you!