Got a question about Student Insight? Here are some of our FAQs …
My institution is not ready for using student predictions, is it possible to set this up later?
Student Insight provides two ways for you to identify students at risk.
Student outcome analytics enables staff to see how student, academic and activity factors have affected student outcomes such as academic performance in the past. Staff may select individual factors and see how it affects a student's likelihood of non-continuation or poor academic performance. Student Insight is able to display students who match these factors and "tag" these students to group them into a risk cohort. Staff may then monitor the progress, engagement and outcomes of students by cohort.
In addition, Student Insight can learn a predictive model from your institution's historical data. This identifies the patterns in the data that affect student outcomes and may be used to generate a risk score that predicts the level of risk that a student may withdraw from their course or fail an individual unit of study.
These may be setup independently, or used together to help staff gauge which students may be at risk.
What data do I need to use Student Insight?
Student Insight allows you to load data into the system from any data source that contains information about the student, their activities and attainment. Data from each source is combined into the system into Feature Sets which represent a particular view of the student and their interactions with the institution. Student Insight will then generate a predictive model for each Feature Set which are combined to generate an overall prediction of risk for the student.
As a minimum, Student Insight needs data about each student enrolled on a course at the institution, and requires the following information:
- Student characteristics and demographics
- Student historical module or unit of study results
- Student summative assessment data
- Student enrolment information such as the programme that the student is taking and their route of study
- Information about each course such as course subject
- Information about each module or unit of study, such as subject and credits.
This data is typically stored in your Student Management System.
Additionally, Student Insight may use activity data stored in logs generated from student interactions with systems such as the VLE and Library. These are used to train predictive models which use student engagement with teaching and learning resources to predict student risk and allow staff to view student engagement patterns.
How much historical data do I need to train Student Insight?
As a minimum at least 1 full academic year of historical data is required from your Student Management System.
Can I change the predictive model?
Many learning analytics solutions provide a fixed model that cannot be changed by the institution. Student Insight is unique in that it allows the predictive model and the data sources it uses to be customised by the institution. Each institution is unique and has different characteristics and challenges. Rather than provide a “one size fits all” approach, Insight provides tools to enable the predictive model to be tailored to best reflect your institution’s situation and students.
How do I know why the predictive model has identified that a student is at risk?
Learning Analytics solutions often generate a single overall prediction of the student’s outcome, but do not allow staff to understand how the prediction was calculated. In contrast, Student Insight allows staff to see what factors have led to an increased risk for the student. Users can view this information on the Influence Chart.
The Influence Chart allows staff to assess which factors have the most influence on the student’s predicted outcome. This can be useful information to help the staff understand the issues the student is facing, and decide what might be appropriate courses of action to take.
How accurate are student predictions?
The accuracy of the predictive models that Student Insight can build depends on a number of factors, such as the amount of historic data available, data quality, cohorts of students included in the model, and the risk metric against which it is targeted. Student Insight provides tools that enable you to assess the accuracy of the model in your context, and optimise it to ensure that the predictions it makes meet your objectives based on the financial cost of interventions and the impact the positive and negative predictions.
As a guide, academic performance predictions based on student demographic, attainment and historical academic performance, have an accuracy of 79%.
Do I need SITS to run Student Insight?
No. Student Insight is a standalone product which can integrate with any Student Management System.
Can students use Student Insight?
Currently, Student Insight provides dashboards that staff may use to monitor student progress and potential outcomes. In addition, subject to security controls, staff may access an individual student record in order to see more information about the student's progress, predicted risk and outcomes. This becomes a useful aid in discussing with the student whether they are facing any issues and planning what the next steps should be.
It is our belief that student analytics of this kind need to be discussed openly as part of a partnership between staff and each student, in order to ensure that they are interpreted positively and used to help inform next steps. As a result, students are currently not able to directly access analytics, however we are working closely with our customers to design student facing dashboards which provide the right, positive information to students which can be interpreted in the right way.
Is it web hosted or do I have to host it from my server?
An institution may choose to host Student Insight in house. Alternative Student Insight is available as a hosted solution, fully managed and supported by Tribal.
What systems does Student Insight integrate with?
Student Insight is highly configurable and allows data to be loaded from any data source that contains data about students and their attainment, attendance or activities.
The system is integrated directly with the Tribal SITS student management system and SID student support desk.
Do I have to supply data in a predefined format?
Student Insight allows data to be loaded in flat file format, such as comma delimited and fixed width files. These may be loaded manually or pushed to the system through integration with the Student Insight API.
Can staff log into the system using the login they use for other systems?
Student Insight supports Single Sign On with the SITS e:Vision web portal. This allows a member of staff who is already logged into e:Vision to log directly into Student Insight.