The Influence of Decaying the Representation of Older Social Media Content on Simulated Hiring Decisions
PhD Study - 2019
Context
Decaying representations (featured in detail in another PhD study on the portfolio) gradually make social media content less visible to readers over time, which can help users disassociate from past online activities. The literature shows that online reputation informs managers' hiring decision, however, existing privacy tools are burdensome, and no usable solutions exist yet.
The study explored whether shrinking, one decaying representation, influences managers’ assessments and simulated hiring decisions of job candidates, compared to seeing a full profile or an empty profile with no posts.
Contributions Overview
The study showed that shrinking or hiding posts resulted in significantly more positive hiring decisions and assessment than the full profile. It also showed that managers contintue to use online reputation in their hiring practices.
The shrinking representation further promoted a positive impression of candidates, suggesting that this type of User Interface representation can be important to online reputation management, especially when social media privacy management tools lack usability.
More Details about the Method and Results
Objective
The study tested the effect of applying shrinking on a fictitious Facebook profile on simulated hiring decision. It also expolored whether the application of shrinking on a female or male candidate’s profile impacts managers’ decisions, and looked into whether managers’ demographics such as gender and age influence their hiring decisions. The study focused on an US context and probed the current use of online reputation by US managers.
Prototype
Six Facebook profiles in total were prototyped and contained fictitious data; three profiles for a fictitious female user and three profiles for a fictitious male user). The three conditions were: original, decayed, and control.
The decayed representation showed a Facebook profile with posts gradually shrinking in size. The original representation showed a Facebook profile with all posts in full size. The control representation showed a Facebook profile with no posts.
The two control profiles showed no posts on the timeline, as if the user had made all posts private. The content of the four prototypes with visible posts was identical except for small changes to match the user’s gender.
Method and Analysis
3×2 between-subjects study with 360 US managers, with type of representation (original, decayed, control) and the candidate’s gender (male, female) as the independent variables.
Responses were collected through an online crowdsourcing survey where participants received a link to a video displaying one profile corresponding to their study condition. Participants were randomly assigned to one of six conditions. After viewing the video, managers were asked questions relating to hiring the candidate and asked them to provide their evaluation of the candidate against several criteria and personality traits.
Results are reported based on statistical analysis of the survery 5-point Likert-scale responses and a qualitative analysis of the survey's open-ended questions.
Videos
All six videos are available is this playlist: https://youtube.com/playlist?list=PLjrjidbvU9kb71hSjUputTN74Zg-fxOsE
Below is a sample video showing the gradually shrunk posts on the fictitious female candidate.
Findings
Decay representations that apply shrinking positively impacted impression and the simulated hiring decisions, except for aspects of online reputation that could be perceived as negative where the Control representation led to a more positive impression. In addition, middle-aged managers minimized negative online activities of the female candidate, while older managers minimized negative online activities of the male candidate, suggesting that stereotypes still exist and can affect the hiring process.
Consistent with the literature, online reputation was found to continue to inform hiring decisions at the time of the study, emphasizing the need for tools that help users easily dissociate from past online content.
Study Details in Depth
The full study details, including the literature review, methods, other prototypes, results, and discussion can be accessed through the published paper.
The paper was published in Proceedings of the 2020 ACM CHI Conference on Human Factors in Computing Systems, which is one of the premier conferences in the field.
