Information Bias and the Importance of Using Data Analytics Responsibly

A data analyst sits at a desk working on their computer.

The field of information management grows more important by the moment. Data expands continually — and with it, the potential to gain valuable insights.

Data analytics keeps supply chains optimized and allows retailers to predict which products a consumer will buy next. It also allows analysts to track the spread of COVID-19 and monitor patient outcomes in different regions across the United States. Information professionals perform mission-critical analysis for their organizations, but as the tools they use to manage and analyze data grow more sophisticated, their responsibility to use those tools ethically becomes even more evident.

Whether detecting fraud for financial institutions or enabling evidence-based care for health-care providers, analysts must uphold sound practices first and foremost. Their responsibilities begin with gaining an understanding of the risks associated with improper data management, which allows them to adopt practices and create standards that support their social responsibilities.

Information Bias and Other Data Management Challenges

Professionals who manage and analyze data have to keep a number of critical issues in mind:

  • Data privacy and consent encompasses an individual’s ability to protect and control the privacy of information about them, as well as an individual’s option to prevent the collection and use of their personal data.
  • Transparency in data collection and usage refers to the responsibility that professionals and organizations that gather personal data have to be open about the collection and use of that data.
  • Information bias describes a prejudice or deviation from truth that arises when data is reported or classified incorrectly, or contains inherent imbalance of categories.

Any failure to handle data responsibly or respect the privacy of information holders can have serious consequences. Violations of privacy can result in personal financial or medical information becoming public and falling into the hands of bad actors, and concealing data collection and usage practices can destroy the trust of customers, patrons, and partners. In many cases, privacy violations also break laws and the social contract with communities.

Information bias typically has a more insidious effect. When information professionals introduce, replicate or amplify an error, even unwittingly, they fail to accomplish their core mission: discovering truth and insight. Such a failure can result from confirmation bias, which occurs when an analysts selects or omits data that doesn’t support a preconceived notion; it can also take the form of selection or availability bias, which occurs when sample data that is selected or made available for a study doesn’t accurately represent a population.

Minor instances of information bias might simply reduce the accuracy of a study’s finding, but information bias can also create far more serious consequences. For example, a study of health outcomes associated with a particular medical procedure or drug might fail to detect risks that are unique to a certain population if that population is not represented in the survey sample.

Information Management Responsibilities

Armed with knowledge of risks associated with data analysis, information professionals can lead efforts to manage data responsibly within their organizations.

Handling Personal and Sensitive Data Responsibly

An organization has a responsibility to keep comprehensive records of the information it has, as well as its collection method, storage location, sensitivity level and associated risks, and protections and access restrictions. Organizations also have a responsibility to collect, store and use only data they need. They should not retain sensitive information beyond the time necessary for its intended use.

Practicing Transparency When Collecting Data

Customer data can inform the development of enhanced services — providing, for example, customized user experiences that take into account specific user preferences — but the processes, policies and reasons for collecting that data should be available to the public and clear to the user, and the user should receive an opportunity to opt out of having their data collected.

Creating Responsible Standards for Data Collection and Use

Internal standards and policies for data collection and usage should go beyond simply meeting requirements of laws such as the European Union’s General Data Protection Regulation (GDPR). Organizational policies can address such issues as the methods for collecting and sharing data, as well as what practices are prohibited. For example, an organization might have the ability to gather both activity data and personal identification information from a website user but choose to collect only the activity data if that fulfills its purposes.

Eliminating Bias in Algorithms and AI Systems

Bias can appear at any point during a data analysis project, from conception to analysis. Therefore, the most important step that information management professionals can take to avoid bias is to acknowledge that it exists and build safeguards and countermeasures into their analyses. Data analytics can be an incredibly powerful tool for turning information into meaningful action, but the field depends on accurate data and socially responsible professionals.

Developing Responsible Information Management Professionals

The world creates data at an astounding rate. Market intelligence firm IDC estimates that data created between 2020 and 2023 will exceed the total amount created during the previous 30 years. The exponential growth of information fuels the demand for experts with the training and knowledge to manage huge data sets, detect patterns in raw data and glean actionable insights.

Data management professionals use such insights to help their organizations gain competitive advantages and make societal improvements. The University of Washington Information School develops innovative leaders with foundational knowledge of information management and strategic planning and analysis skills. It prepares graduates to be data-driven and socially conscious information leaders.

With specializations in business intelligence and data analytics, UW’s Master of Science in Information Management (MSIM) program offered online has degree tracks structured for early- and mid-career students looking to advance their careers and take on leadership roles in their organizations. Visit the program to learn more about the MSIM degree and discover how you can use information as a tool to spur positive social and organizational change.

Sources

Analytics Insights, “Top 5 Biases in Data Science to Know for Model’s Accuracy”

Forbes, “Data Transparency in the Age of Privacy Protection”

Harvard Business Review, “Customer Data: Designing for Transparency and Trust”

Harvard Business Review, “What Do We Do About the Biases in AI?”

HealthITAnalytics, “Big Data Analytics Show COVID-19 Spread, Outcomes by Region”

IBM, Journey to AI Blog, “Optimize Healthcare Delivery and Reduce Costs With Prescriptive Analytics”

International Association of Privacy Professionals, Glossary of Privacy Terms

International Data Corporation, “IDC's Global DataSphere Forecast Shows Continued Steady Growth in the Creation and Consumption of Data”

Storage Networking Industry Association, “What Is Data Privacy?”

TechTarget, “8 Types of Bias in Data Analysis and How to Avoid Them”