by Brian Swider, Brad Harris,Murray Barrick
Job interviews typically begin with a set of seemingly innocuous questions unrelated to the job: How is your day going? Got any plans for the weekend? How was traffic on your way in?
It is commonly assumed that job candidates and interviewers both prefer to start with these types of questions rather than just diving into the more rigid and formal structured interview topics. After all, small talk is typically how most interactions between strangers begin. Interviewers also believe these little interactions, academically referred to as “rapport building,” help to loosen up nervous job candidates and lead to candid responses in the subsequent job-related questioning. (Note: Although this premise is intuitive, research has yet to substantiate it.)
These rapport-building conversations may not be as innocent as they seem. Let’s take the classic structured interview, whereby interview questions are typically related to the job, predetermined, and standardized across all candidates in an attempt to eliminate as much “noise” as possible. Decades of research has established that one structured interview has greater accuracy than more than three unstructured interviews — and more effectively helps to select the best candidates. Nevertheless, interviewers and candidates almost always insist on warming up with quick chit-chat (sometimes before they ever sit down or reach their designated interview space), which begs several important questions: Does the brief, free-flowing nature of rapport building undermine the structured interview? Or is it possible that useful, job-relevant information is conveyed during small talk that actually enhances the validity of the interview?
The latter possibility, while seemingly incongruent with interview research, has some support in other fields of study. Research in social psychology, for instance, robustly indicates that individuals can perceive a decent amount of valid information about people just from observing brief excerpts of their social behaviors. In essence, people make first impressions instinctively; it is a survival mechanism rooted in millennia of evolution. While certainly not perfect, research suggests that these initial impressions (derived as quickly as in five seconds of interaction with a stranger) allow a person to have a rough but nevertheless partially valid idea about another individual’s personality, trustworthiness, and intelligence. In the context of the structured interview, this could mean that otherwise superficial questions about the weather or local sports teams might offer a useful sneak peek at how the candidate will perform on the job.
To investigate this possibility, we conducted and video recorded 163 interviews with job-hungry, and realistically nervous, twenty-somethings in a business school master’s program. During these interviews, we had our trained interviewers quickly note their initial impressions of each candidate immediately following a brief (two to three minutes) rapport-building phase (before the formal questioning began). After rating this small talk, they asked the candidates a series of structured and predetermined interview questions and rated each answer. Next, to parse out any bias from the rapport-building phase and to provide a useful comparison data set, we trimmed the rapport-building conversation off the interview videos and had experts watch and rate the quality of each candidate’s response to each question.
What did we find? First, and consistent with mounds of other research, our comparisons revealed that there were differences between the two data sets, suggesting that the initial impressions did have an effect on interviewers. Second, and more novel, our findings suggested that a significant amount of the variance in interview scores provided by the experts was explained by job-relevant attributes (i.e., not unrelated biasing material) that were conveyed in the initial rapport building phase rated by the interviewers. Indeed, this effect held even after we controlled for other common self-presentation tactics during the interview, indicating that the initial impression rating was detecting more than just a candidate’s tendency to schmooze or self-promote. In this sense, rapport-building questions, even when purposefully crafted to be as innocuous and unrelated to the job as possible (as was the case in our test), do seem to provide interviewers with abbreviated, albeit meaningful, insights into a candidate’s job prospects.
Of course, this evidence alone does not do much to disentangle the extent to which an interviewer’s initial impression is biasing versus informative. To address the underlying process, we looked at how each interview question was rated by interviewers, in sequential order, rather than the overall aggregated score from all questions. This approach allowed us to identify when the initial impression was having the largest influence on interviewer ratings of candidates responses. The results indicated a distinct pattern that was consistent with research on primacy effects. Specifically, the impact of initial impressions was shown to be the strongest on the first structured interview question that was rated, but then steadily diminished on subsequent question ratings. That is, the candidates who made good initial impressions received higher ratings than warranted by their answers on the first couple of questions, while candidates who made poor initial impressions received lower ratings on those questions. But this influence on interviewers wore off over time as interviewers were able to gather more and more job-related information.
Our study offers several useful takeaways for candidates, interviewers, and organizations. Candidates need to be “on” during all interactions with prospective employers, even the initial chit-chat. Interviewers can, and likely will, use this information to make inferences about candidates’ suitability for the job, especially early on in the interview. Candidates need to come off as professionally sociable and competent during the opening small talk with interviewers.
For organizations and interviewers, our findings indicate that rapport building, even when specifically directed toward supposedly nonbiasing topics irrelevant to formal interview questions, can influence interview scores. Yet we caution against just eliminating rapport building from interviews, for several reasons. First, although rapport building can easily be seen as unnecessary noise, the interview is often just as much about recruiting as it is selecting. If candidates are turned off by an organization’s sterile interview procedures that forbid the exchange of pleasantries and small talk, then any possible validity gains made by eliminating unstandardized rapport building would likely be offset by more candidates turning down job offers and choosing to work at a friendlier, kinder organization. Second, our results indicate that at least some job-related information is being transmitted and received during rapport building, suggesting that not all of the variance in interview outcomes is due to harmful bias.
Our findings also suggest some promising middle ground. Organizations could standardize the rapport-building questions asked, formalize initial impressions ratings, or both. Alternatively, they could consider dropping the scores on the first few questions of the structured interview (similar to what is done in polygraph testing), as our results indicated those were the questions most influenced by initial impressions. First impressions will occur no matter what, so you might as well be intentional about trying to account for them.
Source: Havard Business Review