> ## Documentation Index > Fetch the complete documentation index at: https://docs.assesskit.com/llms.txt > Use this file to discover all available pages before exploring further. # Preview > Test your experiment before collecting real data # Preview Preview mode lets you test your experiment exactly as participants will experience it. Run through your entire study to verify timing, check instructions, test response collection, and ensure everything works correctly before sharing with real participants. ## Overview Preview mode provides: * **Participant-identical experience** - See exactly what participants see * **Full functionality testing** - All features work as in real experiment * **Data collection** - Preview responses recorded for verification * **Flow testing** - Verify component sequencing and branching * **Timing verification** - Check that durations feel appropriate * **No permanent data** - Preview data separate from real participant data **When to use Preview:** * After building any new components * Before sharing experiment with participants * When making timing or flow changes * To verify randomization works correctly * To test variable substitution * Before submitting for publication or sharing **Critical:** Always preview your complete experiment before collecting real data. Errors discovered during data collection waste participant time and research resources. ## Starting Preview ### Opening Preview Mode **Three ways to start preview:** 1. **Keyboard shortcut:** Press `P` from any view in Task Editor 2. **Preview button:** Click "Preview" button in top toolbar 3. **Preview panel:** Open Preview panel from left sidebar ### Preview Options When starting preview, you can choose: **Start from beginning:** * Runs complete experiment from first component * Recommended for full testing **Start from specific component:** * Jump to middle of experiment * Useful for testing specific sections * Select component from dropdown **Test subset of trials:** * For experiments with many trials (e.g., 160-trial Stroop) * Preview first 10-20 trials instead of all * Configured in frame or variable settings ## Preview Controls During preview, you have control buttons to manage the test session. ### Control Buttons **Pause/Resume:** * Pause experiment at current screen * Resume to continue from same point * Useful for: Taking notes, checking configuration, discussing with colleagues **Skip Forward:** * Jump to next component * Bypasses current component duration * Useful for: Getting past long delays, testing flow connections **Skip Backward:** * Return to previous component * Re-test specific screens * Useful for: Reviewing components, testing variations **Stop/Exit:** * End preview session immediately * Returns to Task Editor * Useful for: Found issue that needs fixing, testing complete **Restart:** * Start experiment over from beginning * Keeps preview mode open * Useful for: Second pass with fresh perspective, testing randomization ### Keyboard Shortcuts During Preview | Key | Action | | ------- | -------------------------- | | `Space` | Pause/Resume | | `→` | Skip to next component | | `←` | Skip to previous component | | `Esc` | Stop preview and exit | | `R` | Restart from beginning | ## What Preview Shows Preview mode replicates the exact participant experience. ### Participant View **What you see:** * Same screen layout participants will see * Same fonts, colors, images, and styling * Same element positioning * Same component sequencing * Same timing (durations, ITIs, timeouts) **What's different:** * Preview controls visible (can be hidden with `H` key) * Current component name shown in header (optional) * Preview data tag in database (doesn't mix with real data) ### Response Collection **All response methods work:** * Keyboard responses recorded * Button clicks captured * Form inputs collected * Mouse interactions logged * Reaction times measured **Where preview data goes:** * Stored temporarily in preview database * Visible in Analytics panel * Not mixed with real participant data * Can be cleared or reviewed **Check response collection:** 1. Complete preview trial 2. Open Analytics panel 3. Verify response recorded correctly 4. Check reaction time measurements 5. Confirm correct/incorrect scoring ### Timing Accuracy Preview shows real timing: * **Component durations** - Actual millisecond timing * **Response windows** - Timeouts and delays work correctly * **ITIs and fixations** - Inter-trial intervals shown * **Auto-advance** - Components transition at specified times **Test timing by:** * Using stopwatch to verify durations * Noting if pacing feels rushed or slow * Checking if response windows are appropriate * Ensuring fixations are visible long enough ### Flow Transitions Verify experiment flow works correctly: * **Sequential flow** - Components advance in correct order * **Loop iterations** - Loops repeat correct number of times * **Randomization** - Order changes across preview runs * **Conditional branching** - Different paths based on responses * **Frame behavior** - Frames execute as configured ## Testing Checklist Systematically verify all aspects of your experiment during preview. ### Before You Start * [ ] Save experiment (preview uses saved version) * [ ] Review checklist of what to test * [ ] Prepare notepad for issues found * [ ] Clear previous preview data (if needed) ### Instructions and Consent * [ ] **Consent form displays correctly** * All text visible and readable * Consent button works * Required reading time enforced (if applicable) * [ ] **Instructions are clear** * No typos or grammatical errors * Instructions match actual task * Examples shown correctly (if applicable) * All images/diagrams load * [ ] **Navigation works** * "Continue" or "Next" buttons functional * Keyboard shortcuts work (if applicable) * Can't accidentally skip important info ### Stimuli and Content * [ ] **All images load correctly** * No broken image links * Images sized appropriately * Image fit mode works as intended * [ ] **Text displays correctly** * Font size readable * Colors provide adequate contrast * Alignment looks professional * No text overflow or truncation * [ ] **Elements positioned correctly** * Centered elements actually centered * Alignment looks intentional * No overlapping elements * Spacing consistent ### Response Collection * [ ] **Response keys work** * All valid keys register responses * Invalid keys ignored * Keyboard focus working correctly * [ ] **Buttons clickable** * All buttons respond to clicks * Button states change (hover, click) * Correct button advances/submits * [ ] **Forms functional** * Text inputs accept typing * Radio buttons single-select * Checkboxes multi-select * Dropdowns open and select * Required fields enforced * Validation works (email format, number ranges, etc.) ### Timing and Pace * [ ] **Component durations feel right** * Not too fast to read/respond * Not so slow participants get bored * Fixations visible but not excessive * [ ] **Response windows appropriate** * Enough time for thoughtful responses (untimed) * Challenging but achievable (speeded tasks) * Timeouts don't occur prematurely * [ ] **Overall pacing comfortable** * Experiment doesn't feel rushed * Breaks at appropriate intervals (if long experiment) * Can maintain attention throughout ### Flow and Structure * [ ] **Components in correct order** * Logical progression (consent → instructions → task → debrief) * No missing steps * No unexpected jumps * [ ] **Loops work correctly** * Repeat expected number of times * Exit properly after iterations complete * [ ] **Randomization functions** * Order changes across preview runs * All components appear (none skipped) * [ ] **Conditional branching works** * Correct path taken based on responses * All branches reachable and tested ### Data Collection * [ ] **Responses recorded** * Check Analytics panel after preview * All responses captured * Correct/incorrect scored properly * Reaction times reasonable * [ ] **Data structure correct** * Expected columns present * Data types appropriate (numbers, text, etc.) * No missing or null values unexpectedly ## Preview with Timeline Variables For experiments using timeline variables, preview with extra checks. ### Testing Subset of Trials Don't preview all 160 trials - test a representative subset. **How to preview subset:** 1. Before preview, go to Variables view 2. Enable "Preview mode" or "Sample trials" 3. Set sample size (e.g., 10 trials) 4. Run preview with just those trials 5. Verify variable substitution works correctly **What to sample:** * Mix of all conditions (if applicable) * Example: If 4 conditions, preview 2-3 trials of each * Check randomization across sampled trials ### Seeing Variable Substitution Verify variables fill in correctly: **Check substitution for:** * **Text content** - `${word}` becomes actual word * **Images** - `${stimulus}.png` loads correct image * **Colors** - `${color}` applies correct color * **Correct responses** - `${correctKey}` matches stimulus **How to verify:** 1. Note variable values in Variables view for first few trials 2. Run preview 3. Check that each trial shows expected stimulus based on variable values 4. Confirm variation across trials (not all showing same stimulus) ### Checking Randomization If variables are set to randomize: **Test randomization:** 1. Run preview once, note trial order 2. Restart preview 3. Note trial order again 4. Verify order changed 5. Confirm all trials still appear (none lost due to randomization) **Check intermissions:** If using block intermissions (e.g., break every 40 trials): * Preview with enough trials to hit first intermission * Verify break screen appears at correct interval * Check break can be ended (continue button works) ## Common Issues to Check Look for these frequent problems during preview. ### Missing Images **Symptom:** Broken image icon or blank space where image should be **Causes:** * Image not uploaded to media library * Incorrect image filename in properties * Wrong file path **Fix:** 1. Open Screens view 2. Select image element 3. Re-select image from media library 4. Verify image appears in preview ### Unclear Instructions **Symptom:** Confusion about what to do, ambiguous wording **Causes:** * Instructions too brief or too detailed * Jargon or technical terms * Missing examples * Response key mappings not explained **Fix:** 1. Revise instruction text for clarity 2. Add example trial or screenshot 3. Explicitly state response keys 4. Test with colleague unfamiliar with task ### Confusing Response Mappings **Symptom:** Unclear which key does what, mistakes on easy trials **Causes:** * Non-intuitive key assignments (e.g., X for "Yes", C for "No") * No reminder of response mapping during task * Too many response options **Fix:** * Use intuitive keys (F/J for left/right, Y/N for yes/no) * Show reminder on each screen (e.g., "F = Red, J = Green") * Simplify to fewer options if possible ### Timing Too Fast or Slow **Symptom:** Can't keep up / getting bored **Causes:** * Stimulus duration too short * ITI too long * Response window insufficient * No self-pacing allowed **Fix:** * Adjust component durations based on preview experience * Pilot with someone else to get objective feedback * Check literature for standard timing in similar tasks * Consider self-paced option for non-timed tasks ### Components in Wrong Order **Symptom:** Unexpected jumps, illogical flow **Causes:** * Dragged components to wrong position * Connections set up incorrectly * Frame encompasses wrong components **Fix:** 1. Check Timeline view for sequence 2. Check Flow view for connections 3. Reorder components as needed 4. Re-preview to confirm ## Tips for Effective Preview ### Preview Early and Often Don't wait until experiment is complete: * **Preview after adding each section** - Catch issues early * **Preview after major changes** - Verify changes work as intended * **Preview before sharing** - Final quality check **Benefit:** Small issues are easier to fix than discovering 10 problems at once. ### Test on Different Devices If participants will use various devices: * **Desktop** - Standard for lab experiments * **Laptop** - Smaller screen, different keyboard * **Tablet** - Touch interface, mobile layout * **Smartphone** - Small screen, vertical orientation **How:** Use responsive design features and test actual devices, not just browser resize. ### Get Colleague Feedback Ask someone unfamiliar with your experiment to preview: * **Fresh perspective** - See things you've missed * **Naive participant viewpoint** - Identifies unclear instructions * **Actual user testing** - Real responses and confusion points **Ask them:** * What was unclear? * Were instructions sufficient? * Was pacing comfortable? * Any technical issues? ### Run Through Complete Experiment Do at least one full-length preview: * **Experience pacing** - Feel how long experiment actually takes * **Test fatigue effects** - Is it too long? * **Check breaks** - Are breaks placed appropriately? * **Verify ending** - Ensure experiment completes properly **Tip:** Time yourself - if it feels long to you, it will feel long to participants. ### Take Notes During Preview Document issues as you find them: * Keep notepad or document open * Pause preview to write notes * Include component names and specific issues * Prioritize critical vs. minor fixes **Example notes:** * "Instruction component 'Task Overview' - typo in line 2" * "Stimulus duration too fast - increase from 1000ms to 1500ms" * "Missing image in trial 5 - check Variables row 5" ### Check Analytics After Preview Verify data collection: 1. Complete preview 2. Open Analytics panel 3. Check data structure 4. Verify all expected columns present 5. Confirm values make sense 6. Check correct/incorrect scoring **Look for:** * Missing data (should be none) * Unexpected values (e.g., negative RTs) * Scoring errors (correct responses marked incorrect) *** ## Next Steps After preview testing and fixes, you're ready to share: Publish and collect data from participants Configure trial variables if not yet done Get help improving your experiment Adjust experimental structure if needed Preview is your quality control step - use it thoroughly to ensure professional, error-free experiments that provide reliable data and good participant experiences.