Unintended Use
Definition¶
Uses of produced data that are not as intended by the researcher.
Definition source: Combatting Bias definition (based on conversations with partner projects).
Stakes¶
part of: opacity, discrimination
related to: multivocality
Similar to risk assessments: by considering unintended uses during research, some of these uses could already be mitigated. Noting unintended uses also flags responsibility and accountability to other researchers.
Where does it occur in the lifecycle?¶
4 - Analyse
5 - Preserve & Share
Questions to consider throughout your work¶
- Can the visualisation be misinterpreted?
- Can the research data be misinterpreted?
- Can the data be used for a different intended purpose (so-called function creep)?
- Can any information in your dataset be viewed as misleading?
Examples¶
- J. Rosen and D.R. Curtis, “Dangers of Non-Critical Use of Historical Plague Data,” Emerging Infectious Diseases, Vol. 24, Issues 1-6 (2018), pp. 103-110.
Good-better-best practices¶
| Good | Better | Best |
|---|---|---|
| Make clear in your mission statement what the goal of your research (output) is. | Discuss any unintended uses (and intended uses) in documentation (e.g. mission statement, final publications). | Work at mitigating unintended uses during your research. |
| Acknowledge the shortcomings of the tools and methods that you use and assess the implications it can have on output and subsequent usage. | Anticipate unintended use of your dataset by future users and warn against such usage. | |
| Consider how you convey this knowledge and put it to use (by conceptualising and implementing mitigation strategies) in future data projects. |