Learning Analytics Policy

Introduction

The collection and use of data about students and their learning is providing new opportunities for institutions to support learners and to enhance educational processes. Learning Analytics systems present visualisations of student learning activity and provide predictive indicators for attainment. These will be used at the University of Greenwich to assist current students in achieving their study goals, and to help us improve our overall provision of education.

The institution will use Learning Analytics to help meet a student-focused vision where, "all Greenwich students are able to excel in their chosen field and to be confident resilient professionals". We are setting objectives to, "create a seamless supportive environment for our students through integrated academic and pastoral support, learner analytics, and opportunities for peer support", and ensure, "our teaching and learning environment will be fit for the 21^stcentury". These are key elements of the University's Academic and Student Experience Strategy (2018-2022).

The university will ensure that Learning Analytics is deployed for the benefit of students, with complete transparency about the data that is being captured, processed and used. The Statement of Principles for Learning Analytics  will be implemented fully and along with the document identifying the Purpose for Learning Analytics  will be publicised widely. All activities in this area will comply with the institution's Data Protection Policy and the General Data Protection Regulation.

Responsibility

Overall responsibility for Learning Analytics at Greenwich is held by Deputy Vice-Chancellor (DVC) (Academic). Responsibility for relevant areas of activity is allocated as follows:

• The collection of data to be used for learning analytics - Director of Strategic Planning/ Director of Information & Library Services (ILS) with the Director of Strategic Planning specifying the types of data to be used and the Director of ILS arranging automated upload of datasets into the data repository and for data security
• The anonymisation or de-identification of data where appropriate - Director of Strategic Planning
• The analytics processes to be performed on the data, and their purposes – Director of Strategic Planning
• The interventions to be carried out on the basis of the analytics –DVC (Academic) working with the Director of Student & Academic Services (SAS) and the Faculty Directors of Student Experience
• The retention and stewardship of data used for and generated by learning analytics – Director of Strategic Planning and Director of ILS working together
• Implementation of learning analytics transparency including feedback of personalised analytics information to students –DVC (Academic).

Analytics presented to students are intended to help them understand how their learning is progressing, and suggestions may be made as to how they can improve their practices. Students are responsible for assessing how they can best apply any such suggestions to their learning.

Students are informed about how their data will be processed when they agree to the relevant Principal Conditions of Registration and associated Student Privacy Notice at registration. Data will be collected for Learning Analytics in compliance with these documents. This information will state the purpose of analytics and the data that will be used for it. It will also mention the involvement of third parties acting as sub-contractors for processing analytics and the rationale for this.

The data for Learning Analytics comes from a variety of sources, including the student record system and the virtual learning environment. The Student Guide to Learning Analytics will clearly specify:

• The data sources being used for Learning Analytics.
• The specific purposes for which Learning Analytics is being used.
• The metrics used, and how the analytics are produced.
• Who has access to the analytics, and why.
• Guidance on how students can interpret any analytics provided to them.
• The interventions that may be taken on the basis of the analytics.

Students will be asked for their consent for any automated prompts or suggestions to be sent to them, based on the analytics. These may include emails, SMS messages or app notifications.

Learning Analytics is separate from assessment. Metrics derived from data sources used for Learning Analytics will not be used for the purposes of assessment and this is stated in the Student Guide to Learning Analytics and the Principal Conditions of Registration/ Student Privacy Notice.

Confidentiality

Personally identifiable data and analytics on an individual student will be provided only to:

• The student.
• University staff members who require the data to support students in their professional capacity.
• University staff in Planning and Statistics who are working in partnership with the data processors to develop and improve the modelling and to evidence the impact of interventions.
• Third parties who are processing learning analytics data on behalf of the institution. In such circumstances the University will put in place contractual arrangements to ensure that the data is held securely and in compliance with the Data Protection legilation.
• Other individuals or organisations to whom the student gives specific consent.

University IT staff will have access to systems and data in order to maintain proper functioning of systems rather than to access any individual's data.

Sensitive Data

The General Data Protection Regulation defines categories of "sensitive data" such as ethnicity or disability. Any use of such data for learning analytics will be fully justified, and documented in the Student Guide to Learning Analytics and any project initiation document or similar.

Validity

The quality, robustness and validity of the data and analytics processes will be monitored by the University which will use its best endeavours to use Learning Analytics in line with best practice in the sector, for example ensuring that:

• Inaccuracies and gaps in the data are understood and minimised.
• A wide rane of data sources are used with the aim of maximising prediction accuracy.
• Interpretation of analytics findings are informed by people with relevant qualifications and experience. This should help avoid over reliance on single findings, for example.
• Written rational justification is used for the choice of algorithms and metrics used for predictive analytics.
• Learning Analytics is seen in its wider context, and is combined with other data and approaches as appropriate.

Student Access to Personal Data

Mechanisms are being developed to enable students to access their personal data, and the Learning Analytics performed on it, at any time in a meaningful, accessible format. Students have the right to correct any inaccurate personal data held about themselves.

Students will also be able to view any metrics derived from their data, and any labels attached to them, though sometimes they may need to request to do so.

On occasion it may be considered that access to the analytics may have a negative impact on the student's academic progress or wellbeing. This may especially be the case when a student's engagement is less than others in a cohort and they are identified as being "at risk". Protocols will be developed to ensure that access this type of data is managed sensitively and that human-mediated guidance is available to the student. However, if the student requests it, all their personal data and analytics will be made available to them. Requests would be made via their personal tutor.

Interventions

A range of interventions may take place with students. The types of intervention and what they are intended to achieve are documented in the Student Guide to Learning Analytics. These may include:

Prompts or suggestions sent automatically to the student via email, SMS message or mobile app notification (subject to the student's consent)

Staff contacting an individual on the basis of the analytics if it is considered that the student may benefit from additional support

Interventions, whether automated or human-mediated, will normally be recorded. The records will be subject to periodic reviews as to their appropriateness and effectiveness.

Minimising Adverse Impacts

The University recognises that Learning Analytics cannot present a complete picture of a student's learning, and that predictive indicators may not always be fully accurate.

Students will retain autonomy in decision making relating to their learning; the analytics are provided to help inform their own decisions about how and what to learn.

Version 1: Updated May 2018

Approving Authority

Consultation Undertaken

University Secretary's Office, Students Union, Student & Academic Services, Communications and Recruitment, Faculties, Staff and Students

Directorate responsible for policy maintenance and review: Planning and Statistics