Unstructured interviews refer to interview formats that lack a fixed pattern, framework, and procedure. Interviewers can freely ask questions to the interviewees, and for the interviewees, there are no fixed standards for answering.
Previous research points out that making good social judgments often requires ignoring information and relying on simple rules. However, this may not be conducive to accuracy. In other words, humans tend to use “simple rules” to find fixed patterns in behavior. If an interviewee’s answers are inconsistent with the interviewer’s impression, the interviewer can dynamically reshape this impression and may ask follow-up questions until they hear an answer that confirms the impression. In unstructured interviews, interviewers may not ask questions that challenge these impressions, leading to impaired final judgment.
To explore whether interviews dilute judgment and worsen it, the authors conducted an experiment at Carnegie Mellon University. Experiment participants predicted the semester GPA of two unknown students: one prediction was based on biographical data (including academic score records, etc.) and an interview, while the other was based only on biographical data (including academic score records, etc.). The results showed that interviewers did feel they obtained more information and understood the interviewee better during the interview, but the actual data indicated that the GPA predictions were significantly less accurate when an interview was added compared to using only accumulated GPA.
In addition to the vast amount of past evidence showing that unstructured interviews do not provide more effective information, this study directly provides evidence that they impair accuracy.
This reminds us that making judgments based on interview impressions may introduce significant biases. More comprehensive information, such as biographical data (including academic score records, etc.), information provided by long-term acquaintances, academic performance records, work achievement records, etc., should be used to assist in judgment and performance prediction.
Ref:Belief in the unstructured interview: The persistence of an illusion
Photo by Christina @ wocintechchat.com on Unsplash