Tasks Test Sessions

Results

Review one session's outcome: the candidate summary, score breakdowns by section, page, dimension, and tag, and the answer panels where open-ended responses are evaluated.

Updated 2026/07/14

The Results view shows one session's complete outcome — what the candidate scored, how the score decomposes, and every answer they gave. It is both the reading surface for auto-scored results and the working surface for evaluating open-ended answers.

The Summary

The candidate's headline result: overall score, pass/fail verdict and performance level where configured, timing, and status. This is the at-a-glance card — everything below explains it.

Score Breakdowns

The score decomposes along every axis the test defines: by section and page (where in the test points were earned and lost), by dimension (competency profile), and by tag (any structured grouping your questions carry). The breakdowns are where a bare 64% turns into “strong numerically, weak on the case-study section”.

Answer Evaluation

Every question appears with the candidate's answer. Auto-scored answers show their computed result; open-ended answers — text, audio, video, code, files, photos — are evaluated here:

  • Score directly, or through the question's rubric when one is attached.
  • Review AI evaluations — the AI's score, feedback, and transcription (for audio/video) — and confirm or override them.
  • The session score updates as evaluations land; results are complete when every question requiring human judgment has it.
Evaluating many candidates question-by-question? Work from the session list per step, keep the rubric visible, and evaluate the same question across all candidates in sequence — consistency improves dramatically when you judge one question at a time rather than one candidate at a time.
Tip

Evaluating Questions with Scored Dimensions

  • The evaluation dialog shows one score input per scored dimension; a per-dialog toggle switches all rows between percentage and points entry (the test's default evaluation mode seeds new evaluations; existing ones reopen in the mode they were entered in).
  • Under the independent strategy an overall score input appears separately — it is not computed from the dimension rows.
  • The question score stays “waiting evaluation” until every scored dimension has a score; deleting an evaluation clears its dimension scores.
  • Questions with scored dimensions must be evaluated in the new application — the legacy interface shows a notice instead of a score input.