If you haven’t heard, LinkedIn recently announced their launch of skill-based assessments.
The assessments are short, multiple-choice tests that users can take to supposedly validate their knowledge on topics like computer languages and other work-related abilities.
We were excited because the biggest recruiting platform in the world now put an emphasis on assessments. They, like us, know that skill assessments will lead the future of hiring.
However, they don’t work.
Tigran Sloyan, a CodeSignal co-founder & CEO, took a bunch of these LinkedIn assessments and found quite a few faults. He wrote a post on LinkedIn, which got tens of thousands of views and was picked up by The Inquirer.
Here’s what happened.
1. Depth of Assessments
In his experiment, Tigran ensured to take assessments on topics he did not have any expertise in. From C++ to Quickbooks, he took a wide variety of tests…and passed them all.
He now has “verified skills” on these topics that he truly knows nothing about. While we trust Tigran to be our CEO, we wouldn’t necessarily trust him with our accounting!
The primary reason that LinkedIn Assessments don’t work is that their questions lack depth to determine a meaningful level of proficiency. They’re too easy to pass (just by guessing and using common sense) and don’t go deep enough.
2. Recruiters Can’t Trust the Results
If anyone can pass an exam, recruiters can’t actually trust the results.
Beyond trusting the results of the actual assessment, LinkedIn Assessments have no way of knowing if a user cheated on the exam. This cheating could be Googling answers or having another person take the assessment.
Without understanding why LinkedIn Assessments don’t work, we fear that assessments as a whole will be dragged down as a bad way to assess skills.
It’s not that assessment as a whole is at fault, it’s the poor design and administration of an assessment that can lead you astray.
3. Theory Vs. Real-World
Another fault in the LinkedIn Assessments is they test theory knowledge rather than real-world abilities.
When you take your driver’s license test, you have to get behind the wheel and actually show your driving abilities on a real road. It’s an actual assessment of your skills versus the theory-based test people take for their driver’s permit.
LinkedIn Assessments are the driver’s permit of skill-based assessments. They test a lot of theoretical knowledge about topics and all questions are multiple choice. They don’t actually assess anyone’s ability to use software like Quickbooks or how they code.
How to Trust Assessments
In order for us to trust assessments, they must meet certain thresholds.
First, assessments must have a certain level of integrity. That means they must actually measure the abilities they say they do and can’t be cheated.
Second, they must have a level of accuracy and not create false positives, as was the case when Tigran passed all of the exams with flying colors.
Three, assessments can’t just be multiple choice, theory-based questions. They need to include real-world simulations.
LinkedIn Assessments don’t yet meet these thresholds, but it was a step toward the right direction. If you’re looking to assess technical skills, check out CodeSignal to see how we are able to administer certified, accurate, and real-world-like assessments.
This blog post is based on the twenty-sixth episode of the data-driven recruiting podcast hosted by CodeSignal co-founders Sophia Baik and Tigran Sloyan. You can find and listen to this specific episode here or check out the video version embedded above.