Today we’re releasing data from the Checkr platform that sheds light on current risk and the role both recent and past behavior play in determining it. Through an analysis of over 70,000 candidates, who were checked on our platform and had a reportable conviction in their background report from 2010, we found that the rate of actionable criminal information appearing on a background report drops significantly over time. In other words, the longer ago a crime occurred, the less likely there is to be subsequent reportable criminal behavior on a background check.
According to our data, 83.1 percent of individuals who had a criminal conviction in 2010 did not have a criminal conviction in the following year. Fast-forward to seven years after the initial conviction, only 1.5% of this population had a criminal conviction, meaning 98.5 percent of this population did not have a criminal conviction.
Within the first year, there’s a dramatic drop off and only 16.9 percent of people have another criminal conviction the following year. Each subsequent year, that percentage decreases on average by 31.2 percent. These findings are likely due to the fact that those who find jobs after committing a crime, reduce their probability of re-offense.
Reviewing data about more recent re-offenses—or lack thereof—reveals that perhaps companies are putting too much emphasis on crimes from the past. With this new information, they can change the emphasis they place on past crimes, evaluate their current processes, and identify new ways to maintain safety and build trust at their organizations.
You can learn more about this data and how your company can re-evaluate trust and safety at your organization, in our new eBook, Rewriting the Rules of Trust & Safety in the Sharing Economy.
This data was collected from a sample size of over 70k candidates that were screened on Checkr’s platform and had a reportable conviction in their background report from 2010. This data includes both felony and misdemeanor convictions from a diverse sample of candidates across the US. The data was anonymized by removing all personally identifiable information before conducting any analysis.