Examine. An exam that measures itself.
Build a question bank for each program, compose exams from it and run them online. After every sitting, Academara checks whether the exam measures reliably and fairly. So your company's certificate means something.
Question bank
Questions by learning outcome and level, with an explanation of the correct answer. Reviewed first, then put to use.
Online exams
Candidates take the exam in your environment. The result is ready straight away.
Unique, yet equally hard
Every candidate gets their own exam. That helps prevent fraud. Academara checks that every exam is equally hard.
The exam measures itself
Every question, every sitting, every exam is checked for reliability and fairness. You don't promise quality, you show it.
- Reliability of every exam calculated
- A check that every unique exam is equally hard
- Every question checked after every sitting
- Insight into scores by topic
Reliability by the standards
Academara calculates how reliably an exam measures, using the recognized split-half method with the Spearman-Brown correction. It is the method psychometricians use. You test against the standards common in assessment: above 0.80 is good, above 0.70 is sufficient.
Fair draw
A separate exam for each candidate raises the question of whether one exam is harder than another. The total weight is always the same. And Academara also measures whether the exams turn out equally difficult in practice.
A bad question does not stay
Every question in the question bank gets a status after every sitting. You see straight away when a question does not work well. So you fix or replace it, instead of leaving it in the exam for years.
Green
The question works.
Amber
The question needs a look.
Red
Action is needed.
Grey
Not enough data yet.
Exam and training strengthen each other
You see how candidates score on each topic, and how the scores are distributed. If a topic keeps scoring low, you know where the training can be sharper.
Clean measurement
Only first attempts count in the analysis. Resits and test groups do not. That way the figures are not skewed by candidates who have seen the questions before.
The measurements get stronger as more exams are taken. Reliability can be calculated from 10 sittings. The verdict on the draw follows once every question has been taken at least 30 times.
See it with your own branding
We set up a demo with your logo and color, and walk you through it.