Cognitive Bias Lab
Our Cognitive Bias Lab is a browser-based judgement and decision task tool for students, teaching demos, and introductory research. Run classic cognitive bias tasks, compare conditions, and export usable data without reducing the whole thing to a survey that feels like it already knows the answer.
It includes four core tasks in one clean interface: Anchoring Task, Framing Effect Task, Confirmation Bias Task, and Base Rate Neglect Task. Use Free Play to explore the tasks quickly, or switch to Project Mode for participant instructions, practice trials, condition labels, and export-ready results.
Personal use is free. Commercial and institutional use requires a licence which covers all our Labs.
Task Setup
Configure the task for this participant. Keep it tidy. It saves time later when you want to remember what exactly you made them do.
Task Instructions
Show these instructions before the task begins.
A short practice round will begin first. These practice trials are not included in the final results.
Task Complete
Participant: | Condition:
What Cognitive Bias Lab Does
Our Cognitive Bias Lab gives you four classic judgement and decision tasks in one clean browser-based tool. It is designed for students, teaching demos, and introductory research where you want something more structured than improvised classroom examples but much easier to use than specialist software. Use Free Play to explore the tasks quickly, or switch to Project Mode for participant instructions, practice trials, condition labels, and export-ready data.
The Four Tasks
Anchoring Task measures how initial values shape later estimates. Participants judge whether a value is higher or lower than an anchor, then give their own estimate, making it useful for showing how starting points can quietly pull judgement in their direction.
Framing Effect Task measures how equivalent outcomes are judged differently depending on presentation. Participants choose between safe and risky options in gain and loss frames, allowing you to compare how framing shifts decision preferences.
Confirmation Bias Task measures evidence selection. Participants choose which piece of information they would check first when testing a claim, making it useful for showing how people often prefer information that supports what they already suspect.
Base Rate Neglect Task measures how category judgements are shaped by statistical context versus vivid descriptions. Participants choose which category seems more likely overall, allowing you to compare responses driven by the broader base rate with responses pulled by the more psychologically colourful cue.
How to Use It for a Simple Study
Choose one task and keep the design manageable. Select an independent variable such as time pressure, reflection time, distraction, fatigue, group discussion, or prior warning about bias. Keep the testing conditions as consistent as possible, run each participant through the same task, then export the results and compare anchor shift, risky choice rates, confirmation preference, or base-rate consistency depending on the task you used.
How to Interpret Your Results
Cognitive Bias Lab gives you trial-by-trial CSV data, which is much more useful than one dramatic headline score pretending to explain human judgement. The sections below explain what the download contains, what to look at first, and how to read the patterns without immediately deciding you have proven that rationality was a brief historical accident.
Your CSV includes one row per trial. That usually means participant ID, task name, condition label, trial number, stimulus or scenario prompt, response, response time, and task-specific columns such as anchor direction, frame type, evidence type, or base-rate consistency.
In simple terms, each row tells you what the participant was asked, how they responded, and what kind of judgement that response reflected.
Open the CSV in Excel or Google Sheets. The first sensible step is to sort or filter by condition label so you can compare groups or testing conditions without turning the file into a suspiciously dense moral lesson about spreadsheets.
After that, focus on the outcome columns that matter for the task you used. This suite is not really about one general “bias score.” Each task measures something different, so it needs to be read that way.
Anchoring Task: Start by comparing mean estimates after high anchors versus low anchors. The main thing to notice is the anchor shift score. If estimates reliably move upward after high anchors and downward after low anchors, the anchor is doing its quiet little bit of damage.
Framing Effect Task: Look at risky choice rates in gain frames and loss frames. The key comparison is whether participants choose the risky option more often under loss framing than gain framing. That difference is the frame shift score.
Confirmation Bias Task: Look first at how often participants selected confirming evidence versus disconfirming evidence. The main number to notice is the confirmation preference rate, which tells you whether people leaned toward supportive evidence more often than challenging evidence.
Base Rate Neglect Task: Start by comparing base-rate consistent responses with description-consistent responses. The key question is whether participants followed the broader statistical context or the more vivid descriptive cue.
Keep the design simple. If you ran a time pressure versus reflection condition, individual versus paired responding, or warned versus unwarned participants, sort the sheet by condition label and compare the main outcome for the task you used.
Anchoring Task: compare the average estimate after high anchors and low anchors in each condition.
Framing Effect Task: compare risky choice rates in gain and loss frames across conditions.
Confirmation Bias Task: compare how often confirming evidence is chosen first in each condition.
Base Rate Neglect Task: compare base-rate consistency rates across conditions rather than just counting one or two memorable choices.
If a pattern appears repeatedly across several trials, it is worth noticing. If it depends on one odd response buried in the sheet, it probably is not the grand unveiling of human irrationality.
A bias effect in this tool does not mean participants are hopelessly irrational or that one condition has exposed some timeless flaw in human thought. Judgement tasks are influenced by wording, attention, fatigue, motivation, and how carefully participants read the scenarios in the first place.
It is also worth remembering that this is a browser-based teaching and introductory research tool. It is very useful for classroom work, demos, and undergraduate projects, but it is not a tightly controlled behavioural lab environment.
Open the CSV, sort by condition, then compare the main outcome for your chosen task before trying anything more ambitious.
In most cases, that means anchor shift for Anchoring, risky choice by frame for Framing, confirming versus disconfirming selections for Confirmation Bias, and base-rate consistency for Base Rate Neglect.
Example Study Ideas
This tool works well for straightforward introductory designs, including:
time pressure vs. reflection time on Anchoring estimates
gain vs. loss framing effects in risky choice decisions
individual vs. paired responding on Confirmation Bias evidence selection
fatigue or distraction effects on Base Rate Neglect judgements
warning participants about bias before the task to see whether behaviour actually changes
A Quick Note on Precision
Cognitive Bias Lab is a browser-based tool designed for teaching, demos, exploratory work, and introductory student research. It is useful, clean, and far less awkward than trying to build these tasks by hand, but it is not intended to replace specialist experimental software or tightly controlled behavioural lab designs.
Well suited to undergraduate work. For PhD-level precision, that is usually where the methods sections get longer, the equipment gets fussier, and the optimism gets a little more fragile.
