Critical Data Studies A Cross-College Collaboration

Glossary: Raw Data

Raw data, in a technical sense, is defined as “any data object that hasn't undergone thorough data processing” (Techopedia 2019). The issue with this definition, however, is that it implies that since the data hasn't been analyzed, processed, or interpreted in any way, it must be factually accurate and unbiased since there hasn't been any human intervention. Many scholars argue against this reasoning because data does not simply exist, but rather it must be collected and generated, and whichever organization is behind generating this data incorporates its own “discipline… and its own norms and standards for the imagination of data” (Gitelman 3). This reveals that even supposedly “raw” data has still been altered and selectively chosen because the data gatherer already had particular goals, intentions, guidelines, and methods prior to generating the data. Certain influences, such as a company’s restrictions and purposes, will always impact the results of data collection. Duster exemplifies this by discussing how the scientific community was gathering data in the 1980's to deliberately support their agenda and claim that there was a sole linkage between biochemical factors and mental illness (Benjamin 322). Although “raw data” can refer to any type of data that hasn't been processed, it can still be equally as biased and should be evaluated with caution.



“Raw Data.” Techopedia,


Benjamin, Ruha. "Scratch a Theory, You Find a Biography: A Conversation With Troy Duster", Chapter 13, pg. 308-327, in Captivating Technology. (Duke University Press, 2019).


Gitelman, Lisa and Virginia Jackson. (2013). Front Matter. “Raw Data” Is an Oxymoron.”


Student Editors: Sarahy Dueñas, CJ Bruns, Nate Garrison and Vishnu Kamagere. We would like to thank additional student editors who would like to remain anonymous for their contributions.