Active Research Infrastructure – Kim Roberts Lab (Child Memory Lab / BrainWorX)
Wilfrid Laurier University
Link: https://submarinejuice.github.io/Child-Memory-Lab/
This repository contains the full research infrastructure, tooling, and experimental website system used in an active cognitive neuroscience and developmental psychology study conducted in Dr. Kim Roberts’ Child Memory Lab at Wilfrid Laurier University.
The project investigates children’s source monitoring, online credibility evaluation, and memory formation when interacting with real-world web content of varying credibility levels. The system is currently deployed in ongoing child memory research, supporting experimental manipulation, behavioral data collection, and eye-tracking analysis.
This is not a mock or demo project.
This codebase supports live studies with children, adheres to research ethics constraints, and was designed for long-term reproducibility by future research assistants.
- Lab: Child Memory Lab / BrainWorX
- Principal Investigator: Dr. Kim Roberts
- Institution: Wilfrid Laurier University
- Population: Children (approx. ages 8–12)
- Research Focus:
- Source monitoring in children
- Online credibility judgment
- Memory accuracy and misinformation
- Attention allocation and visual scanning behavior
The system integrates experimental website manipulation, lesson-based credibility instruction, and eye-tracking data collection to study how children evaluate and remember online information.
Children increasingly encounter information online, yet:
- They struggle to distinguish credible vs. non-credible sources
- Superficial cues (layout, images, confidence) often override factual indicators
- Traditional lab tasks lack ecological validity
This project addresses these limitations by:
- Using real downloaded websites, not synthetic stimuli
- Systematically manipulating credibility signals at the HTML/CSS level
- Capturing behavioral and attentional data during naturalistic navigation
- Using a custom-built automated webscraper, with Selenium, BeautifulSoup, and PlayWright
Real websites are downloaded and altered to create three controlled credibility tiers:
- High credibility
- Medium credibility
- Low credibility
Manipulations are grounded in a predefined credibility criteria table, including:
- Spelling and grammar quality
- Presence or absence of references
- Publication date visibility
- Author identity and credentials
- Advertisements
- Image quality
- Readability and layout professionalism
Each website contains:
-
Purposefully poorly done design
-
Controlled internal navigation
-
Identical informational content across credibility levels
This ensures experimental control while preserving realism.
The platform supports a multi-phase study design:
- Children evaluate multiple websites (high / medium / low credibility)
- No instruction is provided
Focuses on:
- Spelling and grammar
- Publication date
- References
- Useful information indicators
Adds:
- Images and visual persuasion
- Advertisements
- Readability
- Author identity and credentials
Each lesson uses modified websites aligned with the credibility criteria.
The study integrates Tobii eye-tracking software to capture:
- Gaze duration
- Fixation points
- Visual heatmaps
- Attention to credibility cues (e.g., ads, author names, dates)
Raw outputs are exported as CSV files and later processed for:
- Heatmap generation
- Region-of-interest (ROI) analysis
- Time-based attention metrics
- No identifiable child data is stored in this repository
- All study procedures operate under REB approval
- Eye-tracking data is anonymized at collection
- Websites are hosted locally or offline during experiments
This repository intentionally excludes raw participant data.
A core goal of this project is long-term reproducibility.
The repository includes:
- Step-by-step lab manuals
- Data transfer protocols
- Website modification documentation
- Clear separation of experimental phases
A new research assistant should be able to:
- Understand the study design
- Recreate the websites
- Run experimental sessions independently
- Prepare data for analysis
- HTML / CSS / JavaScript – Website manipulation
- Python – Data preprocessing and analysis
- Tobii SDK / CSV exports – Eye-tracking
- GitHub Pages / Local hosting – Controlled deployment
- Markdown / Obsidian – Documentation and lab manuals
✅ Actively used in ongoing research
This system is currently deployed in live child memory studies within Dr. Kim Roberts’ Child Memory Lab, supporting:
- Participant sessions
- Credibility instruction lessons
- Eye-tracking data collection
- Longitudinal analysis planning
Michelle Chala
Computer Science & Psychology
Computational & Cognitive Neuroscience
Business Minor
Wilfrid Laurier University
Research Developer – Child Memory Lab
AI / Data / Experimental Systems
This project prioritizes:
- Experimental validity over polish
- Documentation over abstraction
- Reproducibility over automation
It is designed to survive handoff, not just demonstration.