What is the goal of the HireVue Assessment Score?
The purpose of the HireVue Assessments Score is to help you find successful candidates, save recruiter time, and reduce time to hire. Our I/O psychologists help you identify the best interview questions to ask for your position. Then HireVue software looks at interview responses for behaviors associated with job success and sorts those interviews the top of your list.
What does the HireVue Assessment Score measure?
The HireVue Assessment score is a percentile ranking.* An interview with a higher score has demonstrated more behaviors and abilities associated with job success than interviews with lower scores. This score enables you to sort interviews with more behaviors associated with job success to the top of your list. Interviews with less evidence of the behaviors and abilities that are associated with job success are sorted to the bottom of your list.
Please note: The percentile ranking is the percentage of interview scores lower than that score. For example, a percentile score of 75% indicates that the interview scored greater than 75% of the interview population.
What data is used for the HireVue Assessment Score?
HireVue uses video interviews and game-based challenges to identify tens of thousands of behaviors, traits, and qualities that signal job performance.
- HireVue Interviews look at the Applicant's verbal responses, which are then transcribed and analyzed using Natural Language Processing which assigns meaning to what they've said. No data from the video portion of the interview is scored. We have over 15 competencies that can be measured by our interview assessments.
- Game Challenges measure cognitive skills, emotional intelligence and/or personality. The cognitive games looks at traits like visuospatial and numerical skills, memory, and pattern recognition; emotional intelligence games look at empathy, influence and collaboration; and our personality games look at a combination of agreeableness, conscientiousness, emotional stability, extraversion, and openness.
What do I do with interviews after they are scored?
Ultimately, the next step in your hiring process is up to you. Once interviews have been scored, now you can sort them. We recommend sorting and reviewing your interviews by HireVue Assessment Score from top tier, to middle tier, to bottom tier.
How does HireVue help minimize bias in the Assessment Score?
We search for and ignore anything that could potentially cause bias against certain demographic groups—age, race, gender, etc. Since we are collecting thousands of data points, removing data that contributes to bias does not impact the overall validity of the model. To ensure every model the HireVue team builds is scientifically validated and avoids this scenario, we follow these steps:
- Audit the incoming data and the performance-based model to identify any bias against certain demographic groups.
- Use the video data to predict demographic group membership and identify the features that are most strongly indicative of the adversely impacted group (if adverse impact is found).
- Remove identification of behaviors or traits that are strongly indicative of the adversely impacted demographic group.
- Re-train the model.
- Repeat this process until bias is mitigated.
Example: We may notice that women are systematically scored lower than men. This may be due to overt bias in the training data or the fact that women were underrepresented in the data set. As features (ie. words or phrases) are removed from the scoring model, or have their contribution to the score reduced, it becomes less and less possible to predict the gender of a candidate and therefore, less possible to mimic the discrimination in the data.
Would potentially successful candidates ever appear in the bottom tier?
No assessment or test is perfectly accurate. We strive to make our assessment as accurate and reliable as possible, and we measure both of these qualities to ensure we meet and exceed scientific and legal standards for accurate measurement. Keeping your models updated to our most current versions is an easy way to maximize accuracy.
We can improve this over time by using data from your candidates and (successful) hires. By combining interviews and job performance data from your company, we can fine-tune the model to understand what success looks like at your organization. This does require a large volume of data to retrain the model, but is an excellent way to boost the accuracy of your scores.
Please note that this doesn't happen automatically and your model isn't updating itself over time. To reduce the risk of bias or the model scoring something that it shouldn't, this process always involves consultation from a HireVue IO Psychology Consultant and/or Customer Success Director.
New interviews may have behaviors that HireVue has not been exposed to or learned about. The previous behaviors identified may not be the only ones that can predict success. The top tier has the highest percentage of potentially successful employees. However, you may find successful employees in the middle or bottom tiers, but that is statistically less likely.
How do Competency Scores work within the HireVue Assessment Score?
Each competency that is measured for your position is scored according to the candidates ability in that area. This is generally based on a single question that specifically measures that competency. The exception is communication, which is scored based on the entire interview.
We score a candidate's ability on a 1-5 scale from novice to expert based on our global norm group.
The competency scores are added together to create a combined score. The combined score is what we use to rank candidates against the cohort and generate the percentile ranking score that we use as the “HireVue Assessment Score.”
Remember that the competency scores are based on an external norm group, while the HireVue Assessment Score is based specifically on your candidates. So it is normal that your candidate with an average HireVue Assessment Score (ie. around 50%) may not have an average competency score of 3. If their score is closer to 4 then your candidates are demonstrating a greater level of competency on average than our norm group, and if closer to 2 then they are demonstrating a lower level of competency.