The HireVue score is a percentile ranking of each candidate's interview responses as compared to all other candidates that have been scored by this same assessment in this account.
Are candidates compared across accounts?
No, candidates are only compared within a single account.
Are candidates compared across models?
No, candidates are only ranked against other candidates who have been assessed with the same model.
Are candidates compared across positions?
If the positions all share the same assessment model, then yes, we compare candidates across all positions sharing the same assessment model when generating their HireVue Score.
Does the HireVue Score get updated over time?
No, once a candidate has taken an interview and/or a game-based assessment, their score does not change. The score represents how well they compare to all candidates who have been assessed up to that point in time on that account, using that assessment.
Why do we not re-score (or re-rank) candidates each time a new interview is completed?
Because decisions are made about whether or not to advance a candidate based on the HireVue Score at that point in time, we can't update or change that score or we would risk overwriting actions that have already been taken.
Imaging we had a candidate Bob, who received a 68% HireVue Score. Based on that score, Bob's interview received a "Gold Star" custom status and was advanced to take a Live interview.
Then imagine the next day five really talented candidates all applied and scored better than Bob. Relative to this new pool, Bob has dropped in the ranking below all five. If he were re-scored, he would now receive a 64%, not sufficient for the "Gold Star" status. We can't un-invite Bob from his Live interview appointment. Re-scoring him would create confusion. So the scores are unchanged. The new candidates who score higher than Bob will also be advanced to "Gold Star" status and all of them will be invited to a Live interview.
How does the HireVue score get calculated for Pre-Built Assessments?
With Pre-Built Assessments (PBAs) we score the candidate responses on each question using the assessment model for that question. For most competencies there is a one-to-one relationship between the competency measured and the question asked. Once we have scored each of the measured competencies, we use these scores to generate the overall HireVue score.
If two candidates get the same competency scores, why do they have different HireVue Scores?
- Time
Because candidates are percentile ranked, even candidates with the same competency scores will likely have different HireVue Scores. This is because each candidate is ranked against their peers at the moment they are scored. This peer group (technically called the norming pool) is the group of candidates they are compared against. On each new interview completion the composition of this pool changes slightly.
In our example above, if Bob received an "Advanced" competency score in cognition and Mary also received an "Advanced" competency score in cognition, but took her interview two days after Bob their HireVue scores would likely not be the same.
Remember in our example, the day after Bob took his interview five really talented candidates all applied and scored better than Bob. This will adjust the pool that future candidates will be compared against. If Mary finished her interview after these candidates, she will likely score lower than Bob.
- Competency Score Ranges
Competency scores are displayed as a range of text labels from "Expert" to "Novice." You can think of these as a series of numbers from one to five with five representing an "Expert" label and one representing a "Novice" label.
But in reality, inside each of the competency tiers there are some candidates who scored on the high side of the range, a 4.9 for example, and those who score on the low side of the range, a 4.1. Both of these candidates will get the same competency score of "Advanced," but will receive a slightly different HireVue score as these raw numbers are used to calculate their final percentile ranking, not the one through five integers or the Novice to Expert labels.