Between-subjects randomization for online studies
Multi-condition designs raise one question everything else hangs on: where does the randomization actually happen? A concrete guide to splitting the work — assign the condition at the study level in Open Lab, randomize the order within it in the lab.js task — and to why Open Lab's minimization-based assignment keeps your groups balanced even at small samples.
20 July 2026
Screening out AI participants: adding AI-agent detection tasks to your online study
A new set of short visual tasks in the paradigm library is built to catch a problem that ordinary attention checks miss — online participants that are actually autonomous AI agents. Here is what the tasks are, why they work, and how to add them to a study as a battery.
14 July 2026
Running a lab.js experiment online, end-to-end
A start-to-finish walkthrough of taking a behavioural experiment online — build the task in lab.js, wrap it as a study in Open Lab, recruit participants, and export clean data to analyse — held together by the Task → Study → Project model.
13 July 2026
Can you trust web-collected data? Bots, inattentive responders, and what paradata can (and can't) tell you
You launch an online study and sessions arrive overnight — but how many are from real, attentive people? A practical, honest look at the two different problems (automated fraud and inattentive humans), the defences researchers actually use, and where process-level quality signals fit.
7 July 2026
Running a study in kiosk mode: what the methods literature says about one shared device, many participants
Between the controlled lab and the open web sits a third setting — one shared station where people take part one after another, in person. Here is what the web-experimental methods literature actually supports about doing it well, and how Open Lab's Box / kiosk mode maps onto it.
1 July 2026