Annette Markham (2012) FABRICATION AS ETHICAL PRACTICE, Information, Communication & Society, 15:3, 334-353, DOI: 10.1080/1369118X.2011.641993

This article focuses on methods of fabrication for protecting the privacy of participants in online contexts. Markham argues that traditional ethical and privacy protections for participants are not adequate for "public, archivable, searchable, and traceable spaces." She proposes ethical fabrication, choosing the term "fabrication" to purposefully push back against conservative research perspectives that view fabrication as misconduct.

Privacy Issues Online

While traditionally researchers and journalists public or recorded words as freely available for analysis, as long as properly and appropriately anonymized, the age of the Internet has made this insufficient for protecting the researched. Data mining might link data to specific identifiable information. Using quotes from blogs or social media is easily searchable and thus traceable back to the original poster. Publishing in obscure journals is no longer sufficient for protecting participants, given the archival nature and global scale of the Internet. There is no consensus on privacy, as some participants may want attribution, while others do not want to be linked to their data. Even knowing their posts are public does not mean people are okay with researchers using them; there are norms and expectations around the "flow of information" and the context in which it was posted. Information posted at one point in time may seem harmless, but may unexpectedly re-arrise and cause harm.


Fabrication involves combining, molding, or arranging data for a particular purpose, in this case, the protection of researched people. Ethical fabrication is not misconduct in the sense that it is neither filled with ill-intent or entirely invented data. It is based heavily in real data with positive ethical intent. It can be instead considered a form of analysis or interpretation.

Types of Ethical Fabrication

Composite Accounts

Involves "selecting representative elements from the data set and composing a new original that is not traceable back to the originals."

Events: A composite of events allows the researcher to combine multiple instances or perspectives to present a "typical" or "ideal-typical" description. This allows a reader to understand a typical scenario, including how it feels to be at that event.

People: The composite of multiple participants or people to protect their identities and describe the phenomenon at hand.

Interactions: Conveying meaningful dialogues about the phenomenon, either from data or from memory.

Processes: Convenying the process of some phenomenon (e.g., the introduction of new people to an online forum) at a high level.

Fictional Narratives

"A narrative is not a set of facts that purports to represent knowledge or ‘truth’. Rather, a narrative aims to portray in a rich and compelling way the problematic nature of life (including research). A narrative is an expression of our lived experience. It is concerned, not with facts, but with plausibility." ---Dawson (2007, p. 84)

Narrative is both a way of presenting one's results and a way of embracing research positionality in shaping the research. Narratives may take on different perspectives:

Realist: Seeks to be a translator and provide an "accurate" narrative description as close as possible to the data. Seeks to maintain "objective distance."

Confessional: Include or focus on the researcher's influence on the study, where decisions are made transparent. This can be mislabeled as "autoethnography" due to the close attention to researcher role.

Impressionist: "Wherein ‘cultural knowledge is slipped to an audience’ in poetic, fragmented, dramatic, or startling ways." Looking for engagement more than accurate "objective" representation of the phenomenon.

Types of Narratives
Layered Accounts: May layer voices of participants, researcher, and theory into a single storyline. Layers function together so that perspectives are separate but build a cohesive argument. For example, one might layer different voices on the same phenomenon.

Drama: A process of "transforming ‘data’ into a drama, with characters, action, and plot" to highlight observations in context.

Conservations and Dialogues: Revising interviews into dialogues like those in novels or screenplays to convey a messy or contextualized conversation. Including details about body movements, speech, or surroundings can lend meaning to analysis.

Lending Authority to Fabrication

To establish credibility and rigor about fabrications, Markham suggests:

  • Identifying and defining steps of interpretaion and fabrication.
  • Provide rich descriptions.
  • Detail the researcher's role in the field and analytical processes.
  • Discuss how researcher experience effects the findings.
  • Describe how data are fabricated "without loss of fidelity."