When Less Interviewers Make Better Decisions
They say that an interviewer makes up their mind after 30 seconds, but do multiple interviewers necessarily come to the same decisions? Recruitment is a key part of any organisation’s strategy, and companies need to consider how many interviewers are necessary to recruit the right people. The aim of a frugal company should be to select the best applicants with the fewest number of interviewers.
Condorcet’s Jury Theorem stipulates that adding additional members to an interview panel should increase the chance of making a correct decision (assuming that they each make the right decision at a better rate than the chance average). This suggests that recruiters are faced with a trade-off between accuracy and cost, and that a large team of interviewers is always better. Hence, two heads are better than one.
The reality however is not that simple. Of note, the sheer pressure of presenting to a room full of interviewers may well be enough to hinder the performance of some of the best candidates, but more than that – larger interview teams are more costly, harder to organise and do not in fact always make better decisions. Fifić and Gigerenzer (2014) recently reported two conditions under which less interviewers may make better choices. Firstly, the addition of further interviewers may lead to free riding. That is, the influence and responsibility that each member has over the final decision diminishes leading them to contribute less effort. To solve this problem, it is advisable to make each individual accountable for their own work or, if this is not possible, to not exceed teams of three.
Secondly, when a company knows who its best interviewer is, adding a second to the panel is likely to have a detrimental effect. In their analysis, Fifić and Gigerenzer (2014) consider cases where two interviewers are homogeneous – they both look for the same cues to make their decisions – and cases when the two are heterogeneous, relying on different cues to form their evaluations. The former scenario results in worse candidate selection because both are attempting to play the game by the same rules, but because the poorer of the two interviewers is less able to spot the best talent indicators he will ultimately bring down the overall team average. If this advice is ignored by the superior, why have a second opinion at all?
The latter case, where the two interviewers use different cues, is more complicated but leads to the same result. While some of the best candidates may be chosen by both, there is likely to be greater variation in the number favoured by only one interviewer. This not only results in increased time spent on the selection process but also leads to a large number of sub-optimal candidates being considered for the job. And with a large pool of evenly favoured but unequally talented people to choose from, chances are a number of bad apples will be picked. Crucially, this number is likely to be higher than if the best interviewer had just made the selection on his own.
In the real world of business, there may not be one expert interviewer who stands out above the rest. When all of the interview team are of a similar standard or the assessment ability of each member is unknown then bigger is better. In particular, three interviewers do a far better job than two whereas adding a fourth, fifth, sixth, etc. yields diminishing returns. There is one exception however. Two interviewers are still unlikely to perform better than one.
In conclusion, it is recommended to use panels of one or three experts rather than spending on additional staff, to avoid adding additional interviewers when the general situation encourages free riding, and to have a team of heterogeneous interviewers if multiple are used. One way to circumvent homogeneity is to not have the best interviewer teach others how to identify candidates and instead let them develop their own styles.