Study

Description of the study

The first goal of the study is to continue Marianne Schmid-Mast and Daniel Gatica-Perez’s research based on an existing online platform for video interviews. This interactive platform collects online video interviews from laptop computers through a web browser and a webcam. For each new uploaded video, subroutines immediately extract nonverbal features, including speaking status and prosody (energy, speaking rate, and pitch). Thus, this platform will be adapted and used to gather video interviews from our participants (recruited from the University of Lausanne and the EPFL) including the collection of qualitative and quantitative data on participants’ experience and used of video interview.

The second goal of the study is to develop a prototype for more complete automatic analyses (using existing algorithms) of participants’ responses to interview questions to design potentially more useful feedback. This feedback will be based on analyses of both verbal (e,g., audio speech stream, wording, narrative response components of past behaviors) and nonverbal (e.g., speech rate, pitch, loudness of voice, nodding) behavior. For instance, previous research from Adrian Bangerter shows that individuals have difficulties producing verbal narrative responses to past-behavior questions. When they do so, they often tend to focus on the initial situation, neglecting to describe what they did or what results were obtained. A simple feedback technique thus consists of prompting them to describe the situation (S), their task (T), the actions (A) they undertook, and the results (R). This is often summarized as the “STAR” acronym in practice. We thus intend to develop algorithms for the automatic detection of these components of a narrative response in order to design and provide a more appropriate feedback to our participants.

The third and last goal of the study is to garner insights into the perception and use of feedback by applicants in the context of video interviews. Because this emerging interview format will likely be increasingly common in the future, there is already a substantial online advice market (e.g., thousands of videos on YouTube) about how to conduct such interviews as an applicant. But there is little scientific knowledge about the usefulness of such advice. Moreover, the user experience is also understudied. Thus, we will gather data on our participants’ experience via a mix of self-reported questionnaires and post-study interviews on how they understood the feedback, how they used it to plan the second interview, or their reactions as to how appropriate the level of detail of the feedback was.