What remains after the voice
An interview avatar hides the participant’s face, yet a head turn, a hand movement and a shift of gaze still carry traces of the person. InterView-C opens 27 German survey interviews to research alongside those traces. Withholding raw audio and removing names produces a quieter copy of the encounter, while the body’s behavior over time remains in the data.[1]
The researchers make the release choices explicit. Participants gave informed consent to recording and sharing de-identified data for research. Person, place and institution names become typed placeholders. Access requires a request, and the agreement prohibits commercial use and re-identification. That establishes a narrower circulation boundary than an unrestricted public download.[1]
The shared clock still connects movement to answers about income, employment and relationship conflict. Looking for identity risk only in the name field misses that connection. The paper acknowledges the possibility of matching someone against another movement sample. Tracking at a median 18.4 hertz misses rapid eye events, but does not eliminate the identifying properties of head and hand motion.[1]
The data’s next destination
Breaking that linkage has a research cost. Keeping the question, spoken answer and movement synchronized lets researchers examine the interaction surrounding a short response. Withholding audio has a cost too: an outside team cannot test a new speech recognizer on those recordings. The privacy–reproducibility trade-off is expressed through concrete decisions about which parts circulate and under what access conditions.[1]
The next researcher’s responsibility therefore extends beyond where the files came from. Movement that shares a clock with sensitive answers remains linked to the same person when reused for a new question. Requested access and a use agreement create a useful boundary in that chain. Their effect depends partly on the recipient’s conduct, because the agreement itself does not erase identifying structure from the data.[1]
InterView-C’s useful contribution is to state residual risk alongside its access safeguards. That disclosure makes it easier to follow which parts of the encounter participants agreed to circulate. The relevant journey is concrete: audio is withheld, transcripts are de-identified, and movement stays aligned with sensitive answers. Participants’ research consent supplies the starting point for understanding the boundaries of that journey.[1]