Risk Perception Estimation Task

Field	Value
Name	Risk Perception Estimation Task
Version	main (0.1.0-dev)
URL / Repository	https://github.com/TaskBeacon/T000036-risk-perception-estimation
Short Description	Chinese health-risk judgment task with 3 scenario levels and 1-7 ordinal ratings
Created By	TaskBeacon build pipeline
Date Updated	2026-04-03
PsyFlow Version	local checkout
PsychoPy Version	2025.1.1
Modality	Behavior
Language	Chinese

1. Task Overview

Participants judge the perceived health risk of brief everyday scenarios. The task uses three scenario levels (low_health_risk, medium_health_risk, high_health_risk) and a discrete 1-7 response scale. There is no objective correctness score and no reward feedback.

The runtime is split into human, QA, and simulation modes. Participant-facing wording is stored in config/*.yaml so the same task logic can be audited and localized without editing trial code.

2. Task Flow

Block-Level Flow

Step	Description
Load Config	Load the mode-specific config and task settings.
Collect Subject Info	Collect subject_id in human mode; inject deterministic IDs in QA/sim.
Setup Runtime	Initialize triggers, window, keyboard, and stimulus bank.
Show Instructions	Present the Chinese instruction screen, with optional instruction voice in human mode.
Generate Conditions	Use built-in BlockUnit.generate_conditions(...) with the three risk labels.
Run Trials	Execute the per-trial risk judgment flow for each condition.
Show Block Break	Display block mean rating and mean RT.
Save Data	Write trial-level CSV output and settings JSON.
Finalize	Emit the end trigger, close the trigger runtime, and quit PsychoPy.

Trial-Level Flow

Step	Description
Fixation	Show a central fixation cross.
Scenario Preview	Show the condition-specific risk vignette without response options.
Rating Response	Show the scenario again with the risk question and 1-7 scale; collect one keypress or timeout.
ITI	Show the fixation cross again before the next trial.

Controller Logic

Feature	Description
Condition Scheduling	Uses PsyFlow BlockUnit.generate_conditions(...) with the configured condition labels.
Determinism	Block seeds come from TaskSettings so QA/sim runs are reproducible.
Adaptive Control	None. This task does not adapt difficulty or reward.

3. Configuration Summary

a. Subject Info

Field	Meaning
subject_id	Numeric participant ID in human mode; QA/sim inject deterministic placeholders.

b. Window Settings

Parameter	Meaning
window.size	Window resolution in pixels.
window.units	PsychoPy coordinate units.
window.bg_color	Background color.
window.fullscreen	Fullscreen toggle.

c. Stimuli

Stimulus ID	Purpose
instruction_text	Chinese task instructions.
fixation	Central fixation cross.
scenario_low / scenario_medium / scenario_high	Condition-specific health-risk scenarios.
rating_prompt	Risk question shown during the response window.
rating_scale	1-7 ordinal rating legend.
block_break	Block summary with mean rating and mean RT.
good_bye	Final summary screen.

d. Timing

Parameter	Meaning
timing.fixation_duration	Duration of the fixation screen.
timing.scenario_preview_duration	Duration of the scenario-only preview screen.
timing.response_window_duration	Duration of the rating response window.
timing.iti_duration	Inter-trial interval duration.

e. Triggers

Parameter	Meaning
exp_onset / exp_end	Experiment start/end.
block_onset / block_end	Block start/end.
fixation_onset	Fixation screen onset.
scenario_preview_onset	Scenario preview onset.
rating_response_onset	Rating screen onset.
rating_response_key	Rating keypress trigger.
rating_response_timeout	Rating window timeout trigger.
iti_onset	ITI onset.

f. Adaptive Controller

Parameter	Meaning
task.conditions	Condition labels scheduled across blocks.
task.condition_weights	Optional weights; null means even scheduling.
task.key_list	Valid rating keys (1-7).
task.seed_mode	Seed mode for reproducible block ordering.

4. Methods (for academic publication)

Participants completed a computerized risk-perception judgment task implemented in PsychoPy/PsyFlow. Each trial presented a brief health-risk scenario, followed by a discrete 1-7 subjective risk rating. Trials were organized into three condition levels corresponding to low, medium, and high perceived risk. The task measured ordinal risk judgments and response latency, with no binary correctness or reward contingency. Trial stimuli, timings, and response mapping were config-defined to support reproducibility, auditability, and localization.