# Screening (https://terac.com/docs/researchers/opportunities/screening)

How to design screening questions and configure auto-grading to qualify the right participants.

Screening determines which applicants are qualified for your opportunity. You can combine demographic filters (handled before participants even see the opportunity) with screening questions (asked during the application flow).

## How Screening Works [#how-screening-works]

Filters → Screening Questions → AI Interview → Qualified

1. **Filters** narrow the audience before they see the opportunity. Only experts matching your demographic, geographic, and professional criteria can view and apply.
2. **Screening questions** are asked during the application flow. They can be auto-graded or manually reviewed.
3. **AI screening interviews** (optional) use your research guide to conduct a voice conversation that evaluates fit.

## Screening Question Types [#screening-question-types]

| Type              | Format                       | Use Case                        |
| ----------------- | ---------------------------- | ------------------------------- |
| **Single select** | One answer from a list       | "What is your primary role?"    |
| **Multi-select**  | Multiple answers from a list | "Which tools do you use daily?" |
| **Yes/No**        | Boolean toggle               | "Do you own a smartphone?"      |

### Designing Good Screening Questions [#designing-good-screening-questions]

> Write screening questions so the qualifying answer is not obvious. If
> participants can guess what you want to hear, the screening loses its value.

* Include plausible distractor answers alongside qualifying ones
* Avoid leading phrasing that hints at the correct response
* Keep questions concise and unambiguous
* Test your questions with a colleague before launching

## Qualification Logic [#qualification-logic]

Each answer option carries a disposition that determines what selecting it does:

| Disposition            | Behavior                                                                                   |
| ---------------------- | ------------------------------------------------------------------------------------------ |
| **Must select one of** | At least one option tagged this way must be selected. The default qualifying rule          |
| **Must select**        | This exact option must be selected. Multi-select only, and every one tagged must be chosen |
| **May select**         | Acceptable; neither qualifies nor disqualifies                                             |
| **Reject**             | Selecting this option disqualifies the participant                                         |
| **Flag for review**    | Diverts an otherwise-qualified applicant to a human decision instead of auto-qualifying    |

> On a single-select question there is only one answer to give, so **Must
> select** and **May select** have no meaning there. Use **Must select one of**
> for the qualifying answers and **Reject** for the disqualifying ones.

### Auto-Grading [#auto-grading]

Terac evaluates answers against these dispositions as soon as they are submitted:

* If the qualifying conditions are met and no **reject** answers are selected, the participant passes
* If any **reject** answer is selected, the participant is screened out immediately
* If a **flag for review** answer is selected, the applicant is routed to a human instead
* Results are instant with no manual review required

### Reviewing Applicants Yourself [#reviewing-applicants-yourself]

Under **Applicant review** in the builder's Audience step or in Settings, choose **Manual review** to review screening responses before an applicant starts work, or **Auto-approve** to let qualified applicants start without your review. Auto-approve hides this step from the pipeline. This is separate from **Terac review**, Terac's internal screening review.

Switching to Auto-approve does not approve applicants already waiting for review or override answers marked **Flag for review**. These applicants remain available in **All** and the **Applicant review** status filter, where you can approve or reject them even though the pipeline step is hidden.

Terac staff use the same verdict controls in the builder and Settings: **When the judge says Approve** offers **Manual review** or **Auto-approve**, and **When the judge says Reject** offers **Manual review** or **Auto-reject**. Enabling Auto-reject in Settings asks for confirmation before rejecting pending applicants.

## Launching Without a Screener [#launching-without-a-screener]

A screener is recommended, because it decides who reaches your paid task. Some studies need none, for example when the task screens participants itself. For those, Terac can let your organization launch opportunities with no screening questions. Ask your Terac contact to turn it on. Until then, an opportunity needs at least one screening question to launch.

Once it is on, the builder keeps recommending a screener but no longer requires one. An opportunity with no screening questions shows **Add a screener for better matches** with the screener questions and on the Review step, with a button that adds one. Clicking **Launch at my price** or **Accept and launch** then asks **Launch without a screener?**. Only **Launch without a screener** launches it; **Add a screener** takes you back to the questions.

Without a screener, everyone who matches your filters goes straight to the task, so the builder makes sure you approve the work before anyone is paid:

* **Needs approval** holds every submission for your review, including on an external link, where a `completed` redirect moves the submission to Work review instead of paying it.
* **Expert approves** is not available. A task already set to it is flagged, and the launch waits until you set it to **Needs approval**. Steps that only confirm something, such as an agreement or joining Slack, keep working as before, as long as one task waits for your review.
* **Applicant review** has to be **Auto-approve**. Holding applicants for review needs screening responses to review, so **Manual review** is not available.

You approve each submission from the Submissions tab, and the participant is paid once you do.

If Terac runs its AI screening interview on your opportunity, applicants are still screened before the task, so launching only asks you to confirm.

The same confirmation applies when you convert a pre-recruiting opportunity with **Convert to Active**. Once an opportunity without a screener is live, **Applicant review** stays on **Auto-approve** in Settings.

To launch through the API, see [Launching Without a Screener](https://terac.com/docs/developers/guides/screening-questions#launching-without-a-screener).

## Quotas [#quotas]

Quotas let you control how many participants you need for specific screening answers. This is useful when you need a balanced sample.

For example, if you need 60 participants total with at least 20 from each of three experience levels:

* Set the total participant cap to 60
* Add a quota for each experience level with a target of 20

Recruitment continues until all quota targets are met and the total cap is reached. Without quotas, recruitment is first-come-first-served.

See the [API documentation](https://terac.com/docs/developers/guides/quotas) for programmatic quota configuration.

## AI Screening Interviews [#ai-screening-interviews]

For opportunities that need deeper qualification, you can attach a research guide with a screening policy. The AI moderator will:

1. Conduct a voice interview following your discussion guide
2. Evaluate responses against your screening policy
3. Auto-qualify or disqualify based on the policy rules

This is particularly useful when qualification depends on nuanced answers that simple multiple-choice questions cannot capture.

See [Research Guides](https://terac.com/docs/researchers/opportunities/research-guides) for how to set up discussion guides and screening policies.

## What's Next? [#whats-next]

- [Configure Filters](https://terac.com/docs/researchers/recruitment/filters): Target participants by demographics and geography
- [Set Up Tasks](https://terac.com/docs/researchers/opportunities/tasks): Define what qualified participants need to complete
- [Review Submissions](https://terac.com/docs/researchers/results/submissions): Manage the review workflow