Where the time goes
Test analysis
p-value and item-rest correlation per question, Cronbach's alpha and measurement error for the whole test.
That varies per test. It depends on the handwriting, the type of question and how precise your mark scheme is. Short answers with a clear mark scheme go better than long, reflective answers. There is no independent research into its accuracy yet, so we don't quote percentages. Our advice: mark the first test yourself as well and compare.
How to start
Setting up a pilot
This is how schools usually go about it. It's a suggestion, not a requirement.
Get in touch: Email schools@graidlens.eu with your school, the number of teachers and the subjects. We agree on what you want to test.
Data processing agreement: We send a draft. Have your data protection officer review it before teachers upload real tests.
Administrator and invitations: We set up the school and appoint an administrator, who invites the teachers.
Start small: Start with a few teachers and tests they also mark themselves. That shows where it works and where it doesn't.
Evaluate: Agree beforehand on what you'll judge it by: time, quality of the suggestions, ease of use. Then decide whether and how to continue.
More on rolling this out across a school is on the for schools.
Try it on your next test.
Create an account, set up a test and let GraidLens mark a few students. The first time, mark them yourself as well and compare.
Create account