Ethical guidelines for data handling and learning analytics in Compleap project

Jan 16, 2019

Compleap services are aiming at creating learner centred digital ecosystem of competence development. This involves gathering and using various data ranging from completed previous and ongoing present education, education application attempts made in the past, work experience and hobbies, language skills to open badges as well as acquired competences and future goals.

This information will be used in Compelap services and by learning analytics functionalities and is planned to be coming automatically from databases as well as from manually user provided data.

Databases in question (e.g. KOSKI) deal with very diverse and sensitive information. For instance, KOSKI has information available on the education accomplished in imprisonment institutions as well as on learning, intellectual or physical disabilities a person may have and governmental assistance a person may have used. Unethical, uncareful data handling or even users’ concerns about the data handling can easily compromise project credibility and success as well as put users at risk for unethical conduct.

Personal information of this type is very sensitive and private and requires adequate handling. For this reason, high priority on privacy and data security must be placed during the time frame of the project and even after project is over. At the moment there is no single perspective on handling ethical issues on learning analytics and there is no united or common framework of ethical standards regarding the use of learning analytics for educational purposes. However, most of the scientific sources agree on the importance of consent and clarity.

To help the project team and developers better understand and adhere to the ethical standards we suggest following eight-point ethical code of practice proposed by Sclater and Bailey (2015). The code consists of responsibility, transparency and consent, privacy, validity, access, enabling positive interventions, minimizing adverse impacts and stewardship of data. Will go through all the points to discuss how have these principles beeen applied in the framework of Compleap project.

 

Not all principles, however, are equally developed. For instance, there is still work to be done concerning the transfer and storage of the data and responsibility for that together with database representatives (KOSKI) and developers (SOLITA). Interventions taken will still have to be evaluated to measure the extent of positive effect it has on the users. Feedback from the users will have to be gathered to see how beneficial and sensible the information presented is to them. Adherence to the ethical standards have to be ensured even after the end of Compleap project. As the project goes in to the second year these issues will have to be investigated and solved.

Compleap aims at providing EU-wide digital services that help individuals in competence development. In the further development of the project pre-existing national and Europe-wide registries could be utilized not only in Finland but also in other partner counties. However, the adherence to the ethical guidelines still has to be preserved as these principle can and should be followed even if other data sources and data gathering methods will be used in Eu partner countries.

We further present the table with the principle descriptions and its’ interpretation and application in the Compleap project.

Table 1. Application of cthical code of practice by Sclater and Bailey (2015) in the framework of Compleap project. 

 

Principle

Decription

Application in Compleap

1.

Responsibility

Responsibility – it must be decided who has an overall responsibility as well as specific aspects of learning analytics, e.g. collection of data and interventions. Student and staff body should be consulted as well.

It is still being negotiated who will have final responsibility of different aspects of transferring and storing user data. Priority is being given to safe and ethical handling of user data.

2.

Transparency and consent

All steps of learning analytics have to be clearly explained and transparent. Consent has to be given for any changes with data use or any new projects.

User consent has to be granted to create user profile. Transparency will be ensured by explaining to the user what kind of data and for what purposes is used. User can also choose what kind of data he/she wants to import and use in user profile and for calculating education recommendation. Limitations and interpretation of education recommendation will be explained to the user in a written form.

 

3.

Privacy

Access to students’ data should be restricted and re-identification from metadata and aggregation of different data sources avoided. Cases there institution may have legal obligation to intervene should be clearly stated.

Privacy is ensured by restricting the access to user data. Strong authentication by a trusted service provider is needed to create and access user profile.

Users who don’t want or can’t use strong identification will be able to modify the profile, but it will not be possible to save it, so user privacy is protected.

It still has to be considered what third parties (e.g. education institutions) if any will have access to the service.

 

4.

Validity

Institutions must ensure that data and analytics processes are of high value, valid and also useful and appropriate.

Data will be used only from trusted data sources (defined here as Finnish registries and some user-entered data). We also try to make data presented useful and appropriate. For this reason, testing with end users will be done to make sure the information is understandable and useful for them.

 

5.

Access

Student must be able to access their data at all times. However, institutions may temporarily withhold data if it may have harmful impact on the student’s academic progress.

Users will be able to access their user profile with data in it at all times in the time frame of the project.

6.

Enabling positive intervention

It should be specified when institutions need to intervene and advise students. Their appropriateness and effectives should be reviewed.

Work that has been done is carefully designed to enable positive impact of learning analytics. Visualisation of competence profile as well as education recommendation will be designed to be of value for the user. Moreover, using the service as a whole is seen as an intervention on user reflection and decision-making process and the appropriateness and effectiveness of that will have to be evaluated in the second part of the project.

 

7.

Minimizing adverse impacts

Institution take steps to ensure that students’ data categorization, labeling, norms and trends do not lead to bias, discriminatory attitudes, non-participation, cheating the system, loss of control in learning or any negative effect on wellbeing or academic performance.

Great care has been taken to prevent any undesired effects of learning analytics. For instance, not including any predictive or profiling functionalities in Compleap services as this is would potentially de-motivate those already struggling users, and possibly increase discriminatory attitudes.

8.

Stewardship of data

Learning analytics will be kept to a minimum to deliver desired result and retained only to specified and clear purposes. It has to be in compliance with existing legislation.

Learning analytics is kept to a minimum in Compleap and unnecessary data gathering is avoided. Data is gathered and presented only to support individual in taking control of her/his learning process. Effort is also made to ensure that whole service works according to Finnish and European legislation.