Design a study
This guide walks through each tab of the Study form (create or edit) and explains what to fill in, why it matters, and where those settings appear for participants.
Overview
The Study form is organized into six tabs:
| Tab | Purpose |
|---|---|
| General | Title, description, status, password, UI language, sharing, moderation, and quotas. |
| Recording | Screen sharing, audio calibration, and interaction telemetry. |
| Pages | All participant-facing screens: welcome, device gate, requirements, consent, permission denied, calibration, participation form, moderation, conclusion, closed page, and consumed link. |
| Tasks | Add, order, and configure behavioral, ancillary, and stimuli tasks. |
| Analysis | Post-session processing: Speech-to-Text and eye-tracking. |
| Diagram | Read-only visual representation of the full participation pipeline. |
Save early, then set the Status to Open (in the General tab) when you are ready to recruit. Use Preview to dry-run the entire flow without recording or uploads.
Owners and Editors can update the study. Viewers and Analysts see a read-only form for review.
AI study assistant
At the top of the Add Study and Edit Study pages, an AI study assistant panel lets you design or change a study by describing what you want in plain language. It is available to owners and editors on plans that include the assistant; on other plans the panel shows an upgrade note instead.
What it can do
- Design a study from a description. For example: "Create a study that starts with a welcome screen, then a demographics questionnaire asking age, education, and handedness."
- Change an existing study. For example: "Add a practice block before the main task" or "Add a consent checkbox confirming the participant is at least 18."
It works on study design only. It never reads or changes participation results, recordings, or logs.
How it works (two confirmations)
- Review before anything changes. When you send a request, the assistant proposes a clear list of changes. Nothing in the form changes until you choose Apply changes. Choosing Discard changes nothing.
- Save stays in your hands. Applying fills in the form and marks it as unsaved, exactly as if you had typed the edits yourself. The study is saved only when you click the normal Save / Update button. You can review, adjust, or use Revert this assistant change before deciding to save.
Tips
- Be specific about counts, durations, and order (for example, "20 congruent and 20 incongruent trials in random order, 500 ms fixation").
- The proposed changes are shown as a short summary; large trial sets appear as a one-line count (for example, "40 Stroop trials generated: 20 congruent, 20 incongruent; randomized") rather than a long list.
- The assistant can write study content (instructions, consent text, questionnaire labels) in the language you ask for.
Describe the study design only. Do not paste participant names, contact details, transcripts, recordings, or result data into the assistant. Custom participation-form fields you accept become part of what the study collects, so make sure any new field is covered by your consent.
General
Title and Description
Shown on researcher listings and during participation.
Status
After the first save, a Status dropdown appears in the General tab.
- Keep Closed while drafting.
- Switch to Open to allow anyone with the link to participate.
- Switch to Restricted to require a single-use invite link. Only participants with a valid, unconsumed invite token can join.
Invite links (restricted studies)
When a study is Restricted (or Open), you can generate single-use invite links from the Invite Links dialog on the study participations page.
- Open the participations page for the study.
- Click Invite Links.
- Set a count (up to 500) and optional labels, then click Generate.
- Copy individual URLs and distribute to participants.
Each link contains a unique token that is consumed after the participant gives consent. Once consumed, the token cannot be reused. Labels attached to tokens are automatically copied to the participation document for filtering and export.
- Open studies accept both the general link and invite links.
- Restricted studies only accept invite links.
- Tokens survive study close/reopen and are not auto-revoked.
- On subscription downgrade, both Open and Restricted studies are auto-closed.
Study password
(Optional) Set a password to require entry before participants can click Start. The welcome screen will show a password prompt. You can customize the prompt and labels in UI translations.
UI language and translations
Choose a base UI language or switch to Custom. Click the configure button to open the "Study user interface language" dialog.
The following labels appear during participation but are not part of any task:
- Start, Next, Stop study, Session
- Camera, Microphone, Screen
- Password labels
- Consent checklist instruction
- All built-in initial-form field labels
- Keyboard keycap labels (Space, Enter, Escape, Delete, Backspace, Tab)
Editing any value sets the language to Custom. You can edit translations using the spreadsheet interface in the dialog; the table expands to show all columns and the dialog width adapts to keep labels readable.
The keyboard keycap labels control what participants see printed on the on-screen key (for example the word "Space" or its French equivalent "Espace") in any task that shows a clickable keyboard hint. Only the six named keys listed above are customizable this way; single letters, digits, and the arrow keys are never translated since they display literally or as universal directional symbols. Each keycap row starts pre-filled with the label that matches your chosen UI language; edit a row if you want different wording for that key.
Sharing
(Owners only, after first save)
Click Share study to invite collaborators by email and assign Viewer, Analyst, or Editor access. If the study has not been shared yet, the section shows "Not shared yet." Collaborators appear in the Dashboard studies table with their access level.
Moderation
Moderation controls whether a researcher joins the live session and how much they direct its flow. There are three modes:
| Mode | What happens |
|---|---|
| Unmoderated (default) | The participant completes the study on their own, with no researcher present. |
| Moderated | A researcher must join the live session before tasks begin. The participant waits in a moderation lobby until the moderator starts the session, then moves through the tasks themselves while the moderator observes. |
| Strict | As in moderated mode, a researcher joins first; in addition, the participant cannot move to the next task without moderator approval. |
Enable moderation in the General tab. The Strict control option turns a moderated study into the strict mode above. You can customize the lobby and connection messages in the Pages tab (Moderation section).
Quotas
Enable quotas to limit participation by completed sessions. Set a total participant limit; once reached, new participants see the closed study page (configured in the Pages tab).
A session counts toward the limit as soon as it finishes, whether the participant completed all tasks, a moderator stopped the session early, or the study time limit was reached. The quota is checked before the participant begins, so once the limit is full no further data is collected.
Pages
Use the left sidebar to jump between sections without leaving the tab:
Welcome
Rich text shown on the welcome screen. You can insert images directly; they are uploaded to Storage.
Device gate
Displayed when a participant's device does not meet the minimum requirements (screen resolution, camera, microphone). Use the rich text editor to customize the courtesy message shown alongside the auto-generated requirements checklist.
The checklist uses technical values only (e.g., screen dimensions, browser name and version) so no translation is needed.
Requirements
Optional rich text describing eligibility requirements, plus a list of checkboxes participants must confirm before continuing (e.g., over 18, browser requirements).
Consent
Rich text consent notice (e.g., recording disclaimers) plus a list of checkboxes (e.g., permission to record). All checks must be ticked to proceed.
Permission denied
Rich text shown when a participant denies (or has denied) one or more required browser permissions, such as camera, microphone, or screen sharing. Use the rich text editor to customize the message that explains how to grant access and retry.
Calibration
Appears when Audio calibration in the Recording tab is set to anything other than None and at least one sentence has been added. The page shows one of the configured sentences, centered on screen, and asks the participant to read it out loud. Background-noise level is measured silently in the first couple of seconds, during which the record button is briefly disabled. After recording, the participant taps a round play button to listen back to the sentence; they can either confirm they heard themselves clearly (which completes the page) or choose I did not hear myself to record another attempt. The system stores the measured voice-onset level used by tasks that need it (such as the Stroop task in voice-controlled auto-advance mode).
If the participant's browser cannot run the audio check, or if no sound is detected from the microphone, the page records a skip reason and participation continues to the next step.
Participation Form
Configure intro text, built-in fields, and custom fields. Reorder, mark as required, and save the form shown before tasks begin.
Moderation
Appears in the sidebar only when Moderation (in the General tab) is set to anything other than unmoderated. Customize the participant-facing labels for the moderation lobby: the moderator card title, the waiting-room title and body, and the connection status messages (not connected, connected, ready).
Conclusion
Rich text shown after the last task. Stored as a Quill Delta so styling is preserved.
Closed status message
What participants see if they open a closed study link.
Consumed link page
Appears in the sidebar only for Open and Restricted studies. This page is shown when a participant opens a single-use invite link that has already been used, has been revoked, or does not match this study. Use the rich text editor to customize the message. A neutral default is provided; add your own contact information (email link, lab website, phone number) directly in the message if you want participants to be able to reach you for a new link.
Pages diagram
A read-only reminder diagram at the bottom of the tab shows the flow:
Welcome → Device gate → Requirements → Consent → Calibration → Participation Form
Gray Task and Analysis blocks stand in for hidden details. The Welcome node has a "Gated" connector leading to the Device gate terminal node, a "Closed" connector leading to the Closed terminal node, and (for Open and Restricted studies) a "Consumed link" connector leading to the Consumed link terminal node.
Tasks (Experimental Design)
Adding tasks
Click Add task to open the picker. Each task shows a description before you add it.
| Category | Examples |
|---|---|
| Behavioral | Stroop, N-back, Flanker, Go/No-Go, BART, Sustained Attention (Vigilance), and more. |
| Ancillary | Instructions page, eye-tracking calibrations, questionnaire. |
| Stimuli | Text, image, grid, sequence, video, audio. |
Task order
Each task has a "Next task" dropdown:
- Leave blank to end the flow.
- Point to another task to set the sequence.
Add tasks in any order, then use the dropdown to define the path.
At least one task must be configured. If a participant opens a study with zero tasks, they see the Closed message.
Inline configuration
Selecting a task opens a compact inline panel (no modals or separate tabs) for:
- Task name and instructions
- Media uploads
- Task-specific parameters (e.g., Stroop font/trials, N-back settings, Go/No-Go ratio, BART parameters, vigilance target rule and timing, calibration timing, drawing reference image, questionnaires, stimuli uploads)
The panel stays open while you edit. Select another task to switch, or use Close or Save to finish.
Quick-added tasks come pre-filled with a small, demonstration-sized set of trials so you can preview the task right away. These counts are for demonstration only. Scale the trial lists up to your intended study size before collecting data.
Stimuli image sequences
Drag to set the image order, optionally randomize per session, and set an auto-advance interval for timer-based progression. The Advance after (ms) interval is set in milliseconds. When participants advance manually instead, a Minimum time before next (ms, manual advance) option holds the Next button until that time has passed on each image.
Notable task options
Some tasks include options that change what participants see during the task. The options below are worth knowing before you run a study.
Sustained Attention (Vigilance): response rule and stimulus mode
The Sustained Attention task presents a timed stream of stimuli and records whether participants tap correctly to targets and withhold to non-targets. Three response-rule options let you cover different vigilance paradigms:
- Respond to target: the participant taps whenever the stimulus matches one of the target items in the configured list. Use for auditory reading paradigms where participants tap on a specific letter.
- Respond to all except: the participant taps on every stimulus EXCEPT the items in the withhold list. The withhold items are the non-targets. Note the inversion: the listed items are what participants hold back on, not what they tap on. Review your trial sheet carefully if you use this rule.
- Respond to sequence: the participant taps only when a specific probe immediately follows a specific cue. The first item of each block is never a target (no valid predecessor).
Stimulus mode controls how the trial sequence is built:
- Sheet: you enter each trial row manually, with columns for the stimulus, its type (target or nontarget), an optional block label, and optional per-row timing overrides. Use this when you supply a fixed or licensed stimulus sequence.
- Generator: the platform builds the trial sequence at runtime from the trial count, target frequency, and non-target stimulus list you configure. This is recommended for protocols with many trials to avoid pasting long lists.
Auditory presentation plays audio clips that you provide once in the task configuration: for each distinct stimulus, record it with your microphone directly in the study form (record, listen, confirm or re-record) or upload an audio file, and preview each clip or the full sequence inline. Every participant hears exactly the same recording, so the auditory stimulus is identical and reproducible across the whole study. The configuration cannot be saved while a stimulus is missing its clip or while a clip lasts longer than that stimulus's presentation window (a clip is never cut off; a shorter clip is followed by silence, and the configured duration always sets the pace). Because each clip is a fixed recording started by the platform, auditory timing is far more consistent than synthesized speech; visual mode remains the most precise choice for strict reaction-time paradigms.
When the modality includes audio, an optional Auditory instruction field lets you write a short rich-text message (with a sound-level indicator) shown to participants while the clips are playing. Leave it empty to show nothing.
Vigilance decrement statistics (block-by-block hit rate and RT) are computed when trials have block labels or when the main stream has at least 20 trials (in which case the stream is split into two halves automatically).
The task ships as a content-free container. No clinical instrument's stimulus set is bundled. Researchers supply their own stimuli (either via the sheet or the generator) and are responsible for any licensing requirements that apply to their materials.
Verbal Learning: presentation instruction
The Verbal Learning task presents a word list for study and captures spoken free recall after a retention interval. In the Presentation section of the task configuration, an optional Presentation instruction field lets you write a short rich-text message shown to participants while the word list is being presented (spoken, shown on screen, or both, depending on the presentation mode). Leave it empty to show nothing.
Token Test: delay before the Next button
The Token Test presents a grid of coloured shapes and asks participants to touch or move the requested pieces on each trial. A Delay before Next appears (ms) option (default 500) holds the Next button for a short moment at the start of each trial, so participants read and consider the instruction before they can advance. Set it to 0 to make the button available immediately. The button that confirms a trial and moves on is labelled Next.
Navon task: level cue
The Navon task presents large letters made up of small letters. Because a mixed block asks participants to report different levels on different trials, the task can display a short on-screen cue above each stimulus -- for example "Global (large letter)" or "Local (small letters)" -- so the participant always knows which level to report.
- Show level cue toggle (on by default): turn this off only if you want participants to infer the target level themselves without any cue.
- Global label and Local label: editable text fields for the cue that participants read. Edit them to match your study language or to use neutral wording.
The hint instructions also describe the cue; update them if you change the label wording.
Image grid: fit mode
When you configure an image grid task, a Fit mode option controls how the grid fills the screen.
- Fit all on screen (default): the entire grid scales to fit within the visible area so participants never need to scroll to see all images.
- Original size: images are shown at their uploaded dimensions; a very large grid may extend below the fold.
The default changed to Fit all on screen. If you have already-built image-grid tasks, they will now render with the grid fitted to the screen instead of at original size. Review your task to confirm the new layout matches your intent. If you prefer the previous behavior, open the task and switch Fit mode to Original size.
Operation span: recall-count cue
The operation span task asks participants to recall a sequence of letters at the end of each trial. An optional Show recall-count cue toggle (off by default) displays a row of neutral placeholder slots -- one slot per letter to recall -- during the recall phase. The slots do not fill as the participant enters letters; they only indicate how many items are expected.
Turn this on if you want to reduce working-memory demands from tracking the required recall count, or turn it off (the default) to preserve the standard unaided paradigm.
Questionnaire: item types and scale labeling
The generic Questionnaire task (in the Ancillary category) lets you compose your own items with optional randomization.
- Item types include scale, text area, numeric, MCQ, multi-select, and dichotomous.
- Scale labeling: Choose Anchor labeling (left/right labels) or Full labeling (a label for every scale value). Full labeling respects Min/Max/Step and shows one input per value.
- Questions accordion: Each question opens in an accordion; only one is open at a time. Click a question header to expand or collapse it.
- Questions are auto-labeled
q1,q2, ... by display order, and options are internally indexed.
Matrix reasoning: building items by hand
The Matrix reasoning task (Reasoning / Fluid Intelligence) is a Raven-style pattern-completion task: each item shows one stimulus image and a set of answer tiles, and the participant clicks the tile that completes the pattern. It ships with no templates and no bundled stimuli, so you build every item yourself.
- Upload images first: Upload your stimulus and tile images. Each upload is added to the image dropdowns used by the item rows.
- One row per item: Lay out items in a sheet with one row each. A row holds the stimulus image, the answer tiles (Option 1 to Option 8), the correct tile, the block, and an optional per-item timeout in milliseconds.
- Mark the correct tile: Set the Correct cell to the tile that completes the pattern. It must be one of that row's options, and a row needs at least two options.
- Practice and time limits: Use the Block column to mark a row as
practiceormain. You can optionally enable a practice block and set a per-block time limit. - At least one main item is required: A study cannot be saved or previewed until there is at least one runnable
mainitem (an empty table or a practice-only table is rejected).
Tasks diagram
The tasks-only diagram beside the list updates as you change "Next task" links so you can verify the flow visually. It starts with Start and End unconnected; links only appear once you set a "Next task." The other diagrams stay locked for reference.
Diagram (Visual design)
Open the Diagram tab to see an auto-arranged view of the entire study:
Welcome → Device gate → Requirements → Consent → Calibration
→ Participation Form → Tasks (in order) → Conclusion → Analysis
The Welcome node has up to four outputs:
| Output | Leads to |
|---|---|
| Next | Normal participation flow |
| Closed | Closed study page |
| Gated | Device gate (requirements not met) |
| Consumed link | Consumed link page (Open and Restricted studies only) |
- Labels (e.g., "Start", "Next", connector names) follow the selected UI language.
- Click Fit view to recenter if you pan or zoom.
- The main diagram, Pages snippet, and Analysis snippet are read-only. Adjust task order or content from the Tasks tab.
Quick reminder diagrams also appear in the Pages tab (pages-only view), Tasks tab (tasks-only with Start/End nodes), and Analysis tab (analysis pipeline view).
Recording
Screen sharing
Toggle "Include screen share" to enable or disable screen sharing. New studies have this enabled by default so screen sharing is requested right after consent alongside camera and microphone. Disable it if the study does not require screen capture.
Audio calibration
Run a short microphone and headphones check at the start of each session. The participant lands on a single screen showing one of the configured sentences (centered), reads it out loud, hears the recording played back, and confirms they heard themselves clearly. Background-noise level is measured silently in the first couple of seconds, during which the record button is briefly disabled. The system uses the measured voice-onset level for tasks that need it (such as the Stroop task in voice-controlled auto-advance mode). Audio calibration is automatically switched away from None and locked whenever at least one task uses voice-controlled auto-advance.
| Setting | Description |
|---|---|
| Audio calibration | Single dropdown. None disables the calibration page entirely. Custom displays an editable table that starts empty so you can supply your own sentences. The remaining values are language presets (English, Spanish, French, German, Italian, Portuguese, Japanese, Korean, Mandarin Chinese, Russian) that pre-fill the table with a published audiometric sentence set for that language. Locked off None when at least one task requires voice-controlled auto-advance. |
| Calibration sentences | Visible whenever the mode is not None. Use Add to insert rows and edit the sentences you want participants to read. Each participant is shown exactly one sentence, picked at random from this list when the calibration page opens and kept the same across re-record attempts — different participants therefore see different sentences, which spreads test coverage across your list. Each row has its own Acquisition length (ms) that caps how long the microphone records for the sentence assigned to that participant. Switching the dropdown to another language preset reseeds the table from that language's source (any edits are lost); switching to Custom clears the table. Leaving the table empty skips the calibration page at participation time. |
Language presets cite the published source for their default sentences: Harvard Sentences (IEEE, 1969) for English, HINT-FR (Vaillancourt et al., 2008) for French, Cárdenas & Marrero (1994) for Spanish, Freiburger Sprachverständlichkeitstest (Hahlbrock, 1953) for German, Bocca & Pellegrini (1950) for Italian, Caldas Aulete clinical adaptations for Portuguese, Mandarin HINT (Wong et al., 2007) for Mandarin Chinese, 67S list (Tsuiki, 1973) for Japanese, K-HINT (Moon et al., 2008) for Korean, and a neutral common-usage sentence for Russian.
Real-time speech recognition for individual tasks (such as Stroop in voice-controlled auto-advance mode) is now configured per task. See the task's own settings.
Interaction telemetry
Enable to record non-content interaction signals (clicks, pointer movement, scrolling, key presses).
Analysis
Pick the analysis algorithms that run after uploads finish. Turn on any combination needed for the study.
| Algorithm | Description | Configuration |
|---|---|---|
| Speech-to-Text | Transcribes the microphone audio. | Pick the expected language per study. |
| Eye-tracking (WebGazer) | Detects fixations and builds heatmaps from gaze samples collected during the session. | Two fixation thresholds (see below). |
Eye-tracking thresholds
When you enable eye-tracking, two extra fields appear:
- Dispersion threshold — how spread out the gaze samples are allowed to be before they stop counting as a single fixation. Expressed as a percentage of the viewport. Lower values are stricter. Default 2%, accepted range 0.1% to 20%.
- Minimum fixation duration (ms) — the shortest window of time that can count as a fixation. Default 100 ms, accepted range 30 to 2000 ms.
If you are unsure, leave the defaults — they correspond to typical webcam eye-tracking research settings.
These settings do not affect the live UI. They control the converter pipeline that runs after each session.
Analysis diagram: The analysis diagram beside the list updates as you toggle algorithms, showing the Interview node feeding only the selected processors.
Saving and opening
- Click Create to save a new study for the first time. The study then opens in its full management view, with a side navigation on the left offering three sections: Design (the form you were just filling in), Participations (the sessions collected for this study), and Sharing (the collaborators who can access it). After this first save, you can switch the Status from Closed to Open.
- Clicking Update while editing an existing study keeps you on the same page with a brief success notice, so you can continue editing without being redirected. The same side navigation (Design, Participations, Sharing) is available whenever you open an existing study to edit it.
- When ready to recruit, copy the participation link from the Dashboard. See Share a participation link.
If a study already has participations, you cannot edit it directly. Use Duplicate project to apply changes to a copy, or remove the participations first and then edit the original study.
If your access level is Viewer or Analyst, the form remains read-only so you can review settings without changing them.
Previewing
Use the Preview button to open the full flow without recording, uploads, or password verification. This is useful for:
- Proofreading participant-facing texts
- Validating task sequencing
- Checking the overall participation experience