Analytics Overview

Analytics turns completed test sessions into performance signal — how a test scores across everyone who took it, and how each question performs across every place it was asked. It's built from finished sessions only, so the numbers describe what candidates actually experienced rather than what's currently configured in the editor.

Three reports, one data source

All analytics screens read from the same underlying dataset, just at different scopes:

• Test Analytics — one row per test: submissions, average/min/max score, standard deviation, and (when opened from a specific test) a pass/fail split and score-frequency histogram.
• Question Analytics — one row per question, across all the tests and tasks it's been used in: how often it was asked, how often it was answered, its average score, and its P-index and D-index.
• Per-Question Analytics — the deep-dive on a single question, opened from the Question Bank editor: the same scores plus a breakdown of which choices or match options candidates actually picked.

Where each report lives

Test Analytics and Question Analytics are both standalone pages under Analytics in the left navigation, and are also embedded in context: open a test in the Test Editor and its Analytics tab shows that one test's numbers; open a task and its Analytics tab shows the questions asked within that task. Per-Question Analytics has no standalone page — it opens from the question's own editor in the Question Bank.

Filtering

Every report can be narrowed by task, test, and date range; Question Analytics additionally filters by folder. Selecting a task narrows the tests you can pick (and vice versa), since the two are linked through the sessions that actually ran. Filters are the main way to compare — for example, filtering Question Analytics down to a single test turns on D-index, a comparison that only makes sense within one test's population (see Question Analytics for why).

In this section

• Test Analytics — how test-level scores, pass/fail, and score distribution are reported.
• Question Analytics — how question-level performance (P-index, D-index) is reported across a folder, test, or task.
• Per-Question Analytics — the single-question deep dive, including choice and match distributions.