Lesson 4.3: Data Quality Assurance

Description: Addresses the common data quality issues identified during training (the “7 sins of data quality”), providing practical solutions and best practices for ensuring high-quality data.

By the end of this Lesson, you will be able to:

• Explain the importance of data quality for the ACQF QCP.
• Define and apply core data quality dimensions (accuracy, completeness, consistency, timeliness, validity, integrity, uniqueness) to qualification data.
• Identify common data quality issues, including the “7 sins of data quality,” and their impact in a qualifications context.
• Describe practical data quality assurance processes, including validation techniques and checklists.
• Understand the principles of continuous quality assurance within the QCP.

---

Core data quality aspects for qualifications data

Data quality refers to the state of completeness, validity, consistency, timeliness, and accuracy that makes data appropriate for a specific use. High-quality data is “fit for its intended uses in operations, decision-making, and planning”. For the African Qualifications and Credentials Platform (QCP), this means that the qualification data it holds must be sufficiently accurate, current, complete, and consistent to reliably support its core functions: facilitating the transparency, comparability, and recognition of qualifications across Africa.

Accuracy:

The degree to which data correctly represents the national data specifications it describes.

QCP example: A qualification listed in QCP with an a correct NQF level, as appearing in the official national registration and an appropriately translated ACQF level, or, the registration end date of a qualification is registered correctly in the QCP.

Completeness:

The extent to which all required and expected data is present. In case of QCP, all mandatory fields defined in the African Learning Model (ALM) for a qualification record should be populated.

QCP example: A qualification record is submitted to QCP, with all mandatory and optional fields filled.

Consistency:

Consistency refers to the absence of contradictions or differences when comparing two or more representations of a thing against a definition, or when data is compared across different datasets or systems. Within QCP, this would mean that the data definitions from the ALM are be applied consistently when preparing the national mapping of qualifications to the ALM.

QCP example: Necessary credits for a specific NQF level are indicated consistently across multiple qualifications.

Timeliness:

The degree to which data is up-to-date and available when it is needed or expected. Information about qualifications must reflect their current status. Changes to existing qualifications (e.g., revised learning outcomes, new accreditation period) ought to be updated in a timely manner, and discontinued or expired qualifications clearly marked or archived.

QCP Example: A qualification whose accreditation expired is also archived in the QCP in a timely manner.

Validity:

The degree to which data conforms to the syntax (format, type, range) of the ALM defined rules.

Data entered into QCP fields must adhere to the ALM’s specified data types (e.g., date, integer, string), formats (e.g., YYYY-MM-DD for dates), and value constraints (e.g., NQF level must be a number between 1 and the maximum defined by ACQF). Values for fields governed by controlled vocabularies must be selected from the predefined list.

QCP example: Text such as the thematic area of a qualifications comes from the ISCED-F 2013 taxonomy of fields of education and training.

Integrity:

The accuracy and consistency of relationships between data elements or records, often across different tables or classes in a database.

QCP example: a qualification record in QCP links to an awarding body or an accreditation body, which is an entity already added to QCP and exists as a valid organisation.

Uniqueness:

Ensuring that there are no duplicate records for the same real-world entity in the database.

QCP Example: The “National Certificate in Plumbing Level 3” from Country B is entered twice into QCP, once with the awarding body listed and once without.

---

Data quality assurance processes for QCP

Ensuring high-quality data in the QCP requires a multi-faceted approach involving processes applied before, during, and after data entry or import.

Data preparation

The first step of quality assurance involves concluding all the necessary preparatory steps before the automatic or manual data import to QCP. This includes examining data from source systems (national registers) and its relation to the target structure (the African Learning Model) to understand its structure, content, quality, and interrelationships. For more details on this process, please visit section 4.2.3.

Data entry checklist (during and after submission of data)

The following Data Entry Checklist is designed to ensure that the qualification data being uploaded to the QCP is complete, accurate, consistent, and aligns with the African Learning Model (ALM).

This checklist guides personnel through a structured process of verification, validation, and quality assurance at each stage of data handling. By following this checklist, teams can avoid common errors, inconsistencies, and misalignments with national or international frameworks.

Please bear in mind that this is a generic proposal for a checklist – you may further tailor it based on the national data structure of your qualification (e.g. to include notes on how the mapping process is done to the ALM), the work organisation within your team (e.g. depending on various roles that you allocate to the team members) and the data storage method that you currently use.

Table: Generic Checklist Template

Step	Description	Status (done / not done)
Phase 1: Qualification mapping and preparation
Review source materials	Ensure that relevant personnel has reviewed and has a good understanding of the relevant ACQF QCP Documents (African Learning Model, Data Collection Template and the Architecture Report, as well as the training materials prepared for the QCP Contact Persons).
Verify alignment	Check how the national qualification structure aligned with the African Learning Model.
Document deviations or issues	Record and provide explanations for any areas where alignment with the model is ambivalent. This documentation could be shared with the QCP project team as well.
Decide on common practices	In case of ambiguous fields or potential issues, prepare a national mapping guide with clear transformation rules and circulate it among the personnel inputting data to ensure consistency.
Assess data readiness	Confirm that the source data is fit for mapping by checking for: Completeness: All mandatory fields (e.g., qualification title, NQF level, awarding authority) are present in the source document. Validity: Data formats in the source are predictable (e.g., consistent date formats). Uniqueness: Check if the qualification (using its national ID) has already been entered into the QCP to avoid duplicates.
Phase 2: Data entry
Apply consistent formatting	Verify that all entries follow consistent formats with the platform requirements (you may check the data collection template for more details):
	Cardinality of fields (number of accepted values)
	Date fields
	Numerical values
	Proper capitalisation
	Max character limits
Verify accuracy against source data	Validate critical information and other qualification details, by cross-referencing with source documents or official databases. Mandatory fields to be checked are:
	Qualification title
	Learning outcome title
	Accreditation title (if information class is added)
	Accreditation type (if information class is added)
	Credit points (if information class is added)
	Credit framework (if information class is added)
	Learning opportunity title (if information class is added)
	Learning opportunity provider (if information class is added)
	Language of instruction (if information class is added)
	Any other information fields completed (e.g. unique identifiers, NQF level etc,)
Ensure semantic and relational integrity	Confirm that all relationships and classifications are correct: Controlled Vocabularies: The correct ISCED-F code for ‘Thematic area’ and ESCO codes for ‘Related skills/occupations’ have been selected. Linked Organisations: The ‘Awarding Body’ and ‘Accrediting Organisation’ selected are the correct, existing entities within the QCP.
Check links and references	Confirm that all attached references, URLs, or linked documents are functional and relevant.
Log challenges	Document any issues encountered during the process and share them with your team.
Phase 3: Quality assurance and publication
Conduct cross-field consistency checks	Check that related fields (e.g., learning outcomes and associated qualification levels) are logically consistent and accurate.
System Validation	The QCP system offers automated validation checks through the QCP system to identify errors or missing fields before submission.
Peer review	Request a secondary review from another administrator or curator to catch any overlooked errors or inconsistencies.
Final review before publication	Once peer-reviewed and corrected, change the editorial status to ‘Under review for release’. Confirm that the qualification appears correctly in the staging/review environment of the designated Virtual Space.
Monitor and address feedback	Review and respond to comments or corrections suggested during the peer review or by the QCP users and the project team.

Best practices for maintaining good data quality

Data quality assurance is not a one-time project but a continuous, iterative cycle of improvement.88 The ACQF T3 Training Unit on Mapping and Standardising Qualifications Data emphasizes this through its section on “Continuous Quality Assurance”.

The key principles and best practices of continuous quality assurance include:

• Regular data audits: review samples of QCP data to identify and correct errors or inconsistencies.
• Version control: implement internal versioning rules to track changes to qualifications and establish archiving/unpublishing procedure.
• Clear update procedures: establish clear, documented procedures for updating existing qualification records and for correcting errors discovered post-entry.
• User training and guidance: comprehensive training for all QCP users involved in data entry, mapping, and management is essential to instil good data quality practices.
• Data governance framework: a strong data governance framework (as discussed in SM 3.4) provides the overarching policies and responsibilities for data quality.
• Scheduled reviews: establishing concrete timelines and procedures for systematic reviews) is recommended to maintain the timeliness, relevance and accuracy of data
• Feedback mechanisms: establishing feedback channels for users to report potential errors are vital for identifying discrepancies and improving data reliability.

---

Common data quality issues: the 7 sins of data quality

These “sins” represent common problematic practices and mindsets that can compromise the quality and utility of data within a system like the ACQF QCP. Recognising and avoiding them is crucial for all involved in managing qualifications data:

1. The description dumper: the description dumper copies all properties to the description field and doesn’t map their data to the properties. This practice results in large blocks of unstructured text instead of rich, structured data that the QCP can process. For example, key information (like NQF level, credits, awarding body) becomes buried and loses its semantic value, making it extremely difficult to search, compare qualifications, or perform any automated analysis. It fundamentally undermines interoperability and the comparability goals of the QCP.
2. The learning outcome avoider: this mistake can come in many shapes and forms. Usually, it is either about not introducing any learning outcomes at all or jumbling multiple learning outcomes into just one textbox.  To bring an example of the drawbacks of this practice, not filling in learning outcomes prevent comparing qualifications meaningfully, assessing equivalence, or facilitating credit transfer.
3. The data model hacker: uses ‘other’ and ‘more details’ fields to avoid matching the data model or uses properties and fields for purposes which they are not intended. Data stored in incorrect places bypasses standardisation efforts, making it unreliable for consistent querying, reporting, or linking with other data. This leads to semantic inconsistencies and hinders the platform’s ability to function as a coherent system.
4. The know it all: does not quality assure before posting, assumes data is correct without any basis. This attitude leads directly to the propagation of errors within the QCP. Inaccurate, unvalidated, or outdated information can mislead students, employers, educational institutions, and policymakers and might erode trust in both the national data contributions and the continental platform itself.
5. The dummy data provider: uses dummy date (e.g ‘lorem ipsum dolor…”) to fill required fields temporarily, forgets removing test data from the dataset. The presence of placeholder or test data can lead to incorrect search results, skewed analytics, and general confusion for users trying to find accurate qualification information. It also indicates a lack of rigorous data management and quality assurance processes.
6. The worshiper of old gods: prefers using old data formats and does not plan to migrate. The QCP relies on standardised data models (ALM) and formats for interoperability. Resistance to adopting these new standards means a country’s qualification data cannot be effectively integrated, compared, or utilised within the continental platform. This limits the visibility and international recognition of those qualifications and hinders the overall goals of the ACQF.
7. The minimalist: populated only the required data fields and/or introduced limited information per data field.  While technically fulfilling minimum requirements, providing only sparse information significantly reduces the richness, usefulness, and value of the qualification data in the QCP. It limits the ability for users to make detailed comparisons, understand the full scope of a qualification (e.g. if learning outcomes are too brief), or make well-informed decisions. Thus, the overall utility of the QCP for all stakeholders is diminished.