One of the most prominent trends we're seeing with our customer base is the move to administer more tests remotely, towards the very beginning of the hiring process. As I've talked about previously, there are many advantages to administering tests offsite, mostly to do with the cost and time savings of using objective, reliable data to help you filter out unqualified applicants from large applicant pools. But one question that customers often raise is, "How can I trust test results from remote tests, when I can't verify the identity of the applicant?" Well, unless you're using a remote proctoring service (either human or web-cam based) you can't be 100% sure a candidate took the test without outside help. But what we've found in working with customers is that if your messaging to candidates is well constructed, and you retest candidates onsite, you don't need to worry about the reliability of offsite test results.
We recently did a study with one of our largest customers, who administers aptitude tests remotely at the front of their hiring process, and retests candidates later onsite, if they make it that far. What we found by looking at the data is that the likely percentage of people who didn't take the test honestly (i.e., without outside help) offsite is actually very small, likely much less than 2% of the applicant pool. One reason this may have been the case is that when the employer sent candidates the invitation to take the test, it explicitly described its retesting policy. We really recommend employers do this, because it cuts down on any incentive applicants may have to cheat in the first place. If candidates will have to take a different version of the test onsite anyhow, they are only wasting their own time if they don't take the test on their own the first time.
So a couple of important takeaways. Offsite cheating is not really a big problem, at least according to our data; and you can cut down on it further through proper messaging and by retesting onsite, with those candidates who make it that far.