Sharing iOBE Data Safely to Protect Data Privacy

The iOBE system generates valuable data analytics into student performance, offering deeper insights to educators, students, and institutions alike.

While these benefits are clear, the way in which iOBE data is shared is equally important. All sharing of the iOBE outputs in this website strives to follow relevant data privacy standards.

In this post, I will outline the principles and practices I follow when sharing iOBE data, providing context, examples, and illustrations of how these data are presented and shared safely and effectively.

1. Data is Presented Only in Aggregated and Collective Form

All iOBE outputs that I share online or publish formally are aggregated summaries of assessment results. These are typically presented as Box Plots or Population Charts (as shown in the figure below) that represent the overall performance of a class or cohort.

Key characteristics of this approach:

• No row-level data is released. That is, no dataset is published where each row corresponds to a specific student’s marks across assessment elements.
• Individual scores are never displayed directly; instead, they are merged into collective statistics that summarise group trends.
• No personal identifiers (names, IDs, email addresses, or any other identifiable attributes) are ever included in the published outputs.

This method ensures that while the general performance of the group is clearly communicated, no single student’s results can be isolated or traced back to them.

2. Use of Simulated or Adapted Datasets When Necessary

Where possible, iOBE visuals are generated from simulated or adapted datasets — particularly in demonstration, training, or public outreach contexts.

When real data is used, the following safeguards apply when I want to share those data online:

• Removal of all identifiers before any analysis or visualisation takes place.
• Presentation in aggregated visual form only, never as individual rows or lists.
• Exclusion of small group data where small sample sizes could increase the likelihood of indirect identification.

In certain cases, real datasets may undergo minor controlled modifications (e.g., small randomised adjustments to scores) to further reduce the risk of re-identification (to ensure the privacy of students' data), while keeping overall trends intact.

3. Compliance with Legal and Ethical Data Protection Standards

The iOBE data-sharing process has been designed to align with:

• Best practices in educational data protection, as recommended in academic and institutional guidelines.

Specific compliance measures include:

• Anonymisation by aggregation – Once scores are summarised in statistical form, they no longer constitute personal data under PDPA definitions, provided they cannot be used to identify an individual.
• Removal of identifiers – No names, student IDs, or indirect identifiers are included in shared outputs.
• No re-identification risk – Outputs are constructed in a way that prevents linking back to an individual, even if additional external information is available.

This compliance framework ensures that published visuals are not only useful but also legally sound and ethically responsible.

4. Advantages of Aggregated Visual Sharing Over Row-Based Lists

It is still common practice in some settings to share results in row-based lists—often anonymised using passcodes or partial student IDs. While this may appear to protect privacy, it has several drawbacks:

• Re-identification risk remains – Students can often deduce each other’s codes, especially in small cohorts or close-knit groups.
• Limited insight – Row lists only show individual scores without revealing the broader performance trends.
• Increased complexity for interpretation – Students must scan through multiple rows to guess where they stand compared to peers.

By contrast, iOBE visual outputs:

• Provide instant performance context through statistical summaries.
• Eliminate direct identification risk by avoiding row-level disclosures entirely.
• Communicate trends, variability, and benchmarks more efficiently than lists.

In short, iOBE’s method not only meets privacy standards but also delivers richer and more actionable insights.

5. Purpose and Benefits of the iOBE Data-Sharing Approach

The aim of sharing iOBE data is twofold:

• Educational value – To provide clear and meaningful feedback to students and educators, enabling targeted improvement.
• Trust and transparency – To demonstrate that data is handled with care, in line with legal and ethical obligations.

For educators, these visuals:

• Highlight class-wide strengths and weaknesses.
• Support evidence-based teaching decisions.
• Serve as a communication tool that is both privacy-conscious and information-rich.

For students, aggregated visuals help them:

• Understand their position in relation to class medians, quartiles, and performance ranges.
• Recognise areas where improvement is needed, without the discomfort of being individually exposed.

6. Summary of Data Privacy Safeguards

To reiterate, all iOBE data sharing follows these principles:

• Aggregated data only – No individual records are published.
• No student identifiers – Neither direct nor indirect identifiers are included.
• Effective anonymisation – Data is presented in a form that prevents re-identification.

Final Note

The iOBE data-sharing framework reflects a commitment to maximising educational benefit while minimising privacy risks. By replacing traditional row-based result lists with aggregated visual summaries, educators can share important performance insights in a format that is clear and secure.

For further details on implementing this approach, please feel free to contact me here.

Read more...