Recruitment Screener for UX Research

When to use it

When you recruit from the general public or a broad list, and you need to be sure the person who shows up is the one you need. Interviews, usability tests, focus groups, diary studies.

You do not need one if your audience is small, known and easy to reach — an internal system, a guerrilla test. There a long screener is friction with no return.

Before you use it

This is not the form, it is what goes inside it. The screener gets built in Google Forms or similar; what you download is the question bank, the discard logic and the quota table. Assembling it is the second step and it is described at the end.

Filter on behaviour, not demographics. What someone did predicts better than their age or postcode. Instead of "are you between 25 and 40?", ask "when did you last do X?". Demographics are for distributing quotas, not for deciding who gets in.

Mask the criterion. If it is obvious which answer you want, you will get it — from people who want the incentive. That is why every filter question carries distractor options that filter nothing and exist so the question reads as demographic.

Twenty questions maximum, or five to ten minutes. The screener is not the interview. If you need more, you have too many criteria.

Ask for contact details last, and only from people who passed. Less data stored, less to delete later. And it is what lets you honestly say the answers of people who did not qualify get deleted.

Everything in brackets gets replaced. The worked example lives on this page and not in the downloadable file.

The document

Study: [study name] Profile you are after: [in one sentence, in behavioural terms] Sessions to fill: [10] · over-recruit [20%] Incentive: [amount] per [60] min session Lead: [name · role]

Structure and destinations

The disqualifying questions go first: nobody answers twelve questions only to be screened out afterwards.

Section 1 · Welcome and consent

Opening text, without revealing what you are looking for:

We're looking for people for [a (60) minute conversation over video call] about [general topic]. You don't need to know anything about technology or about [domain]: we're interested in your experience as it is.

This form takes about [4] minutes. If you match the study we'll write to you within [3] working days.

Q1. Do you agree to us using your answers to assess whether you match this study? · multiple choice, required

• Yes, I agree
• I do not agree → section 5

Below it, visible: We keep your answers only for recruiting this study. If you are not selected, we delete them after [30] days. We do not share your details with third parties.

Section 2 · Screen-out

Q2. [Sector or industry question, to exclude anyone too close to the problem] · checkboxes

Q3. [Behavioural question: what they did, in what timeframe] · checkboxes

Never name the client's product here: it conditions everything they answer afterwards.

Q4. [Question that assigns a quota: what exactly happened] · multiple choice, single answer

All four neutrally worded: none of them should sound like the right answer.

Section 3 · Profile and quotas

These do not screen anyone out. They only sort.

Q9 decides the most. A three-word answer, or one copied from the prompt, tells you the person is filling in forms in bulk. It does not screen anyone out on its own: a person reads it.

Section 4 · Contact and availability

• Q10. First and last name · short answer
• Q11. Email and phone · short answer with format validation — "we use them only to arrange the session"
• Q12. When suits you? · checkboxes, at least two ticked

Confirmation message: We've received your answer. If you match the study we'll write to you within [3] working days; if not, we won't contact you. Saying so avoids the round of emails asking.

Section 5 · Closing without selection

No fields. Just the message:

Thank you for taking the time. For this particular study we're looking for a different profile, so we won't continue with the remaining questions.

If you'd like us to let you know about future studies, write to us at [contact]. Your answers to this form are deleted after [30] days.

Quota table

It lives in a separate spreadsheet, fed by the responses. Review it on every recruiting day, not at the end.

Quotas C onwards are minimums met within A and B; they are not added on top. If A and B close with a minimum unmet, you replace a case already booked; you do not add an extra session.

Building it in the form

1. Five sections in the order of the table. In each one, an explicit destination: never leave "continue to next section" in sections 1, 2 and 5.
2. Branching by answer in Q1 and Q4.
3. Q2 and Q3 are checkbox questions, and Google Forms offers neither branching nor content validation for that type — only "select at least / at most / exactly N". The screen-out for those two is reviewed by hand before booking. They stay as checkboxes on purpose: that is what allows the masking with decoys.
4. Randomize the option order in Q2 and Q3, with "none of the above" pinned last.
5. Every question with a visible label and its own help text. The placeholder is not a label: it disappears as you type. Required fields are marked with a word, not only an asterisk.
6. Responses linked to a spreadsheet. The quota table goes on a separate sheet: never write into the responses sheet, the form overwrites it.
7. Before publishing, answer it three times: once passing, once screened out in section 2, once screened out at Q4. If any of them ends in the wrong section, the logic is wrong.

The worked version

The same document with a case written end to end: an invented study on the sign-up flow of a personal finance app. It shows what level of detail each field expects before you face the empty version. It downloads separately.

Internal version. It contains the screen-out criteria, which options are decoys, and the quota table. It is never shared with applicants. What the applicant sees is in the file Screener (participant).

1. Project details

2. Structure and branching logic

Five sections. The screen-outs go first: nobody answers twelve questions only to be left out afterwards. Each section declares where it sends people.

3. Questions with their criteria

The options marked DECOY exist so that nobody can guess what we are looking for. They are not used to filter and they are not analyzed.

Q1 Do you agree to us using your answers to assess whether you match this study?

Multiple choice · required · Yes / No

Criterion: "No" → Section 5. Without consent the answer is not processed and the contact details are not stored.

Q2 Which of these sectors do you work in, or have you worked in over the last 2 years?

Checkboxes · required · tick all that apply

Criterion: ticking banking/fintech or design/UX → Section 5. The six sectors make the question read as demographic rather than as a filter.

Q3 In the last 3 months, which of these apps did you download to your phone?

Checkboxes · required

Criterion: without the finance option ticked → Section 5. The client's app is never named: that conditions everything they answer afterwards.

Q4 With that spending or savings app, what happened?

Multiple choice · required · single answer

Criterion: this is the question that assigns the quota. The first two options go on to Section 3; the other two go to Section 5. All four neutrally worded: none of them sounds like "the right answer".

Q5 – Q9 Profile and quotas

Section 3 · they do not screen anyone out, they only assign quotas

Q9 decides the most. A three-word answer, or one copied from the prompt, tells you the person is filling in forms in bulk. It does not screen anyone out automatically: a person reads it.

Q10 – Q12 Contact and availability

Section 4 · only for people who passed the screen-out

• Q10 — First and last name · short answer
• Q11 — Email and phone · short answer with format validation
• Q12 — Availability slots · checkboxes, at least two ticked

Contact details are asked at the end and only on this branch. Anyone who is left out never hands over their email: fewer details stored, less to delete later.

4. Quota table

It lives in a separate spreadsheet, fed by the form responses. Review it on every recruiting day, not at the end.

Quotas C to F are minimums met within A and B; they are not added on top. If A and B close with a minimum unmet, you replace a case already booked; you do not add an eleventh session.

5. Building it in Google Forms

1. Five sections in the order of the table in point 2. In each one: After section → an explicit destination. Never leave "Continue to next section" in sections 1, 2 and 5.
2. Branching per question: in Q1, Q3 and Q4 use Go to section based on answer. Q2 is a checkbox question and Forms does not allow branching from a checkbox: it is handled with response validation ("doesn't contain" banking/fintech or design/UX) plus a manual review before booking.
3. Randomize the option order in Q2 and Q3 (Shuffle option order). The "None of the above" options stay pinned at the end.
4. Every question with a visible label and its own help text. Not relying on colour or on the placeholder: the tab order has to walk question by question in the order they are read.
5. Theme: accent colour #1d3648, Inter typeface, header with the client's logo area (image 1600 × 400, with alt text).
6. Section 5 with no fields: only the closing message. Do not ask for contact details from someone who is being left out.
7. Responses → Link to Sheets. In that spreadsheet, a separate sheet called Quotas with the table from point 4. Never write into the responses sheet: Forms overwrites it.
8. Confirmation message after submitting: "We've received your answer. If you match the study we'll write to you within [X] working days; if not, we won't contact you." Saying so avoids the round of emails asking.
9. Before publishing, answer the form three times: once passing, once screened out at Q2, once screened out at Q4. If any of them ends in the wrong section, the logic is wrong.

Method note and limitations

The screener does not measure anything: it only decides who gets in. The decoy options reduce strategic answering, they do not eliminate it. The abandonment condition is self-reported and retrospective; if the client has product data, cross-check it before booking. The answers of people who are left out are deleted after [30] days.

UXR SpA · User research [contacto@uxr.cl] · [uxr.cl]

Template v1.0 · reviewed [DD/MM/YYYY] Original at UXR — check that your copy has not expired

The participant version

What the other person receives: the conversation script or the form, with no method notes, no screening criteria and none of the team's hypotheses. It is written over the same example case, so its main use is seeing where the line falls between what gets shared and what does not. It downloads separately.

[Form title: a direct question, not "Screener"]

We're looking for people for [a (60) minute conversation over video call / a (60) minute session]. [One sentence that lowers the barrier: what you do NOT need to know.]

This form takes about [4] minutes. If you match the study, we'll write to you within [3] working days. Everyone who takes part receives [amount] as a thank you.

Section 1 of [4] · Before we start

Do you agree to us using your answers to assess whether you match this study? (required)

We keep your answers only for recruiting this study. If you are not selected, we delete them after [30] days. We never share your details with third parties.

• Yes, I agree
• I do not agree

Section 2 of [4] · About you

The screen-out questions go here, and they never announce which answer screens you out.

[Screen-out question 1] (required)

[Tick all that apply. / Choose one.]

• [option]
• [decoy option]
• [option that screens out]
• [option]
• None of the above

[Screen-out question 2] (required)

• [option]
• [option]
• [option]
• I don't remember

Section 3 of [4] · Your context

The quota questions go here: they do not screen anyone out, they sort.

[Quota question 1] (required)

• [option]
• [option]
• [option]
• Prefer not to say

[Quota question 2] (required)

• [range]
• [range]
• [range]

Section [4] of [4] · So we can contact you

Contact details are only asked of people who passed the screen-out.

Name, or what you'd like us to call you (required)

[ ]

Email or phone (required)

[ ]

Which days and times suit you? (required)

• [slot]
• [slot]
• [slot]

Closing message

For anyone who did not pass the screen-out. No fields: you do not ask for details from someone who is being left out.

Thank you for taking the time. For this particular study we're looking for a different profile, so we won't continue with the remaining questions.

If you'd like us to let you know about future studies, write to us at [contact]. Your answers to this form are deleted after [30] days.

A worked example

A study on the onboarding of a personal finance app. Looking for people who downloaded such an app and did not complete activation in the last 60 days.

P2 · sector. The options are health, education, banking or fintech, retail, design or UX, construction, none. Two discard: banking or fintech — too close to the problem — and design or UX, which answers as a professional rather than as a user. The other four are distractors: they make the question read as demographic.

P3 · behaviour. "In the last 3 months, which of these apps did you download?" — delivery, one for tracking spending or savings, transport, exercise, none. Only the second is a requirement. The client's app is never named.

P4 · quota. "With that spending app, what happened?" — I did not finish creating the account (quota A), I created it and used it (quota B, contrast), I deleted it without opening it (discard), I do not remember (discard).

P9 · quality. "Tell us in your own words what happened the last time you tried to create the account." Two or three lines is enough. Someone who writes "nothing" or repeats the prompt is probably filling in forms in bulk.

This example lives here only, not in the downloadable file.

Check before using it

• Are the discards before the profile questions?
• Does every filter question have distractors, or is it obvious what you want?
• Are you filtering on behaviour rather than demographics?
• Is it twenty questions or fewer?
• Are contact details asked only in the branch that passed?
• Did you answer it three times, with all three outcomes?
• Any bracket left unreplaced?

Grounding

The quality criteria — exact questions instead of "occasionally", avoiding false positives, participants who can express themselves — and the common mistakes come from this site's Screener entry.

The consent in P1 covers the use of recruitment data, which is separate from consent for the session itself: that is the informed consent template. And the screener is the first instrument the research plan lists.