There is a sliding scale based on the value and effort required to complete for each question type. For example, Simulations, Code Challenges, Documents, File Uploads, and Spreadsheets are worth a maximum of 10 points, while simple questions (like Multiple Choice) are only worth 3 points.
Question-type weightings will help to improve the way we grade your candidates' responses and, therefore, help you to continue hiring with confidence.
How will this update help my company hire the best candidate?
Over the years, our question types have improved and increased; with some now being more immersive and requiring a lot more effort from the candidate to complete, which also makes them more insightful. In order to recognize this differentiation between question types, we have introduced question weighting, which will make some question types more valuable to a score and others less. This should help you find better candidates as those that score well on highly-weighted questions will naturally rank higher in the Select stage.
How Questions are Weighted:
Please find an outline of the new weightings in the image below:
- Manual grading always overrides the AI score so candidates that have been manually graded will not recalculate
Example.
An assessment has 6 questions (3 x simulations, 1 x video, 1 x text, and 1 x MC) across 3 different skill groups. In the old weighting example below, each question is worth 10 points. The candidate scored 8/10 across each question, which gives them an overall score of 80%
With the new question weighting in place, if the same candidate has earned the same percentage of points per question (80% of the maximum points, per question), they will still have the same overall percentage of 80%. Although questions are weighted differently, the overall percentage earned is what's calculated in the overall score. This means that questions with less of a value do not automatically drop the overall score:
However, if the same candidate performed better on the simulation questions but worse on the other question types, their score would improve to 84.49%:
And lastly if the same candidate performed worse on the simulations but better on the other question types, their score would be lower overall because the simulations have been weighted to be more important.