Data governance: Data quality

This sub-guideline is part of the guideline Data governance. Refer to the main guideline for context and an overview.

This sub-guideline explains why data quality matters, explores the dimensions of data quality, and examines both the benefits of high-quality data and the risks associated with low-quality data.

Data quality refers to certain features of data that make it accessible, useful and reliable to support effective decision-making. Data must be demonstrably accurate, complete, consistent, accessible, relevant and secure.

High-quality data is essential for parliamentary processes. It forms the foundation of reliable, data-driven decisions that can improve operational efficiency, reduce both operational and strategic risks, and help meet overall business needs. When AI systems are integrated into decision-making processes, they can amplify the benefits of digital solutions. However, it is important to note that this integration can also magnify any associated risks. Therefore, ensuring data quality becomes even more critical when AI is involved in parliamentary operations.

Data quality can be evaluated according to a number of criteria, such as the following:

• Accessibility is the extent to which information is available, or easily and quickly retrievable. Data availability requirements per business process should be identified.

• Appropriate amount of information is the extent to which the volume of information is appropriate for the task at hand. This can be managed by defining, for each data element, how critical the amount of information captured is for analysis purposes.

• Completeness is the extent to which information is not missing and is of sufficient breadth and depth for the task at hand. This aspect is managed by reporting on the completeness of data fields and devising plans to capture all data as per business requirements.

• Consistency is the extent to which information is presented with the same content across multiple systems and platforms. This aspect is managed by understanding data standards and business rules and ensuring that systems adhere to the defined rules.

• Freedom from errors is the extent to which information is correct and reliable, and the degree of agreement between a data value (or set of values) and a source assumed to be correct. Freedom from errors may be attained by defining validation rules, conducting regular testing and reporting on samples of data for compliance.

• Relevance is the extent to which information is pertinent to business processes. This aspect is managed by determining the business use of each data element and assessing its value and relevance through user feedback.

High-quality data has numerous benefits for parliaments, including the following:

• Trustworthy AI systems: Accurate and reliable data enables better-trained AI models.

• Improved analytics: High-quality data enhances the accuracy of analytics, leading to more precise and actionable insights, conclusions and predictions.

• Improved decision-making: Accurate and reliable data allows parliaments to make better-informed decisions with less risk.

• Reduced risk: Secure and protected data can help to prevent fraud, financial losses and reputational damage.

• Cost savings: High-quality data reduces the costs associated with correcting errors, and reduces the time needed to deal with data-related issues.

• Informed policymaking: Accurate and reliable data enables lawmakers to draft bills based on solid evidence, leading to more effective and impactful legislation.

• Support for legislative research: High-quality data is crucial for conducting thorough legislative research, enabling MPs to understand complex issues and make informed decisions.

• Support for innovation: High-quality data can be used to develop new products and services, since the data can reveal not only hidden problems but also potential ways to solve such problems.

• Enhanced public trust: Transparent and high-quality data fosters trust among citizens, who can see that decisions are based on accurate information.

The list below outlines some of the primary obstacles to high-quality data:

• Data integrity issues: Errors in data entry or processing can compromise the accuracy and reliability of data.

• Duplicate data: Having multiple records for the same entity can lead to confusion and errors in reporting and analysis, making it difficult to keep different versions of data in sync across operations.

• Inconsistent data: Variations in data formats or standards across different systems can cause integration issues and inaccuracies.

• Outdated data: Using old or obsolete data can lead to decision-making based on irrelevant information.

• Incomplete data: Missing data fields can lead to gaps in analysis and hinder comprehensive decision-making.

• Ambiguous data: Data that is unclear or lacks context can be misinterpreted, leading to incorrect conclusions.

• Misinformed decisions: Inaccurate or incomplete data can lead decision makers to make the wrong choices. 

In parliamentary contexts, low-quality data can cause a unique set of problems, some of which are discussed below:

• Misguided legislation: When flawed data is used for evidence, research or policy analysis, this can lead to legislation that is ineffective or has unintended negative consequences. For example, an environmental protection bill based on inaccurate pollution statistics might target the wrong industries or fail to address the most pressing issues.

• Ineffective resource allocation: Budgetary decisions based on unreliable data can lead to inefficient resource allocation. For example, allocating funds to a social programme based on outdated poverty statistics might result in the programme not reaching the communities most in need.

• Reduced public trust: The perception that parliamentary decisions are based on questionable data can erode trust in the legislative process and undermine the effectiveness of government institutions.

• Data privacy concerns: Mishandling sensitive data can lead to breaches and loss of public trust. For example, a leak of citizens’ personal data can cause a public outcry and lead to legal repercussions.

• Biases in data collection: Inherent biases in data-collection methods can skew policy outcomes. For example, surveys that do not adequately represent minority groups can lead to policies that overlook their needs.

• Overreliance on quantitative data: Ignoring qualitative data leads to incomplete analysis. For example, focusing solely on statistical data without considering public opinion or sentiment can result in unpopular and ineffective policies and mean that key social trends or attitudes are overlooked.

For example: if a parliamentary committee uses flawed crime statistics to justify increased police funding, the subsequent revelation of data errors could damage public confidence in both the decision-making process and the resulting policy changes.

The Guidelines for AI in parliaments are published by the IPU in collaboration with the Parliamentary Data Science Hub in the IPU’s Centre for Innovation in Parliament. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International licence. It may be freely shared and reused with acknowledgement of the IPU. For more information about the IPU’s work on artificial intelligence, please visit www.ipu.org/AI or contact [email protected].