We’ve featured several employers on the show with recruiting strategies focusing on hiring people with criminal histories. All of these employers have spoken about the benefits for the people they hire, the benefits they get from reaching new talent pools and the benefits for society as a whole.
Despite this growing list of case studies, several misconceptions about hiring people who have been involved with the criminal justice system are holding many employers back.
My guest this week is Shawn Bushway, a professor at the University at Albany and a researcher at RAND Corporation. Shawn has built a body of research on the impact of criminal background checks on employment, and his work informed the current standards of the EEOC. He now helps employers revise their background check policies to better reflect the science about re-offending and become compliant with EEOC guidance.
In the interview, we discuss:
• How background checks have evolved
• Ban the box
• How the tightness of the labour market dictates employers’ attitudes
• Common misconceptions
• The myth around recidivism
• Overestimating the level of risk
• Why crime type is irrelevant
• Making factually informed decisions
• Avoiding false positives
• Age, time and number
• The role of government in creating practical, useable incentives to hire people with criminal histories
Listen to this podcast on Apple Podcasts.
Transcription:
Matt: Support for this podcast is provided by SHL. From talent acquisition to talent management, SHL solutions provide your organization with the power and scale to build your business with the skilled, motivated, and energized workforce you need. SHL takes the guesswork out of growing a talented team by helping you match the right people to the right moments with simplicity and speed. They equip recruiters and leaders with people insights at an organization, team, and individual level, accelerating growth, decision making, talent mobility, and inspiring an inclusive culture, to build a future where businesses thrive because their people thrive, visit shl.com to learn more.
[Recruiting Future theme]
Matt: Hi there. This is Matt Alder. Welcome to episode 548 of The Recruiting Future Podcast. We featured several employers on the show with recruiting strategies focusing on hiring people with criminal histories. All of these employers have spoken about the benefits for the people they hire, the benefits they get from reaching new talent pools, and the benefits for society as a whole. Despite this growing list of case studies, several misconceptions about hiring people who’ve been involved with the criminal justice system are holding many employers back.
My guest this week is Shawn Bushway, a professor at the University at Albany and a researcher at RAND Corporation. Shawn has built a body of research on the impact of criminal background checks on employment, and his work informed the current standards of the EEOC. He now helps employers revise their background check policies to better reflect the science about re-offending and become compliant with EEOC guidance.
Hi Shawn, and welcome to the podcast.
Shawn: Thanks for having me.
Matt: An absolute pleasure to have you on the show. Please, could you introduce yourself and tell everyone what you do?
Shawn: Hi, I’m Shawn Bushway. I’m a professor at the University at Albany and SUNY. I’m also affiliated researcher with RAND Corporation.
Matt: Fantastic stuff. Tell us a bit more about the work you do and what you specialize in.
Shawn: Well, way back when I did my dissertation in 1996, I started looking at the relationship between employment and crime. It was a hot topic at the time, but most of the attention was on the issue that maybe we needed to give people jobs in order to reduce crime. But there was this possibility that it worked the other way too, that if people were involved in crime in the criminal justice system, they would have trouble getting a job. That was a hypothesis at the time that wasn’t well established. I did some of my work in my dissertation, was among the first to show that in fact was true, that people involved in a criminal justice system would then subsequently have problems later.
Now, it promptly was told that I should probably find another topic because no one cared. But as we all know, since around 1999-2000, the topic of re-entry and the need to reintegrate people into society after they’ve been involved in a criminal justice system has become a very big topic. So, most of my career actually has been spent on some element of this issue starting around 2001-2002, where I was recognized as one of the leading experts on background checks, which sounded impressive except for the fact that I was one of the only experts on background checks. But there’s been a lot of rule changes in the United States around background checks because of the changes with the EEOC, Equal Employment Opportunity Commission, which revised its guidance from 1987, finally in 2012, based in part of my work and I’ve been involved in numerous situations with large companies trying to revise their guidance to reflect the new guidance. Then, I went to RAND finally in 2019 for three years to try to see if we could institute or create a broader awareness of some of these issues instead of just doing it one company at a time.
Matt: What do you feel the current awareness is? What’s their attitude or their policy around hiring people with criminal histories?
Shawn: Well, prior to 2012, you could simply say the person has a record, I’m not hiring them. In the United States that was changed and said you have to differentiate between people with records, you can’t just have what they called a blanket band. I think that there were plenty of people hiring those with records. But what happened starting around maybe say 2000 is that the ability to do a background check became a lot easier because of the presence of consumer reporting agencies that worked through the internet and collected criminal histories through the internet for your non-US listeners, America has a fairly unique situation in that conviction, even arrest records are public. And so virtually anyone can find out a conviction record if they would like to, which is obviously a very different situation than in most of the rest of the world.
So, what happened was consumer reporting agencies, these groups that would go out and collect and then sell the information to the companies who then were responsible for making the decision, went from a situation where most companies didn’t to the current situation where most companies do. So, you can no longer just say, “You have a record, we’re not going to hire you.” But most companies now do some form of a background check. There has been a lot more advocacy in that situation because of the increased use. So, we have rules like Ban the Box, which is only true in some states, but basically a lot of large companies have just said we’re going to follow Ban the Box anyways no matter where we are because you can’t have multiple policies.
Ban the Box works the following way. It says you can’t ask about criminal history record on the first application. So, what happens instead is– but it doesn’t say you can’t ask. So, what it does is, you have a situation where people are offered a job contingent on the background check. Then there’s a background check. And the way the EEOC says you should do this is you should have a first-round check just based on the criminal history record. You should be able to clear a fair number of people to make their hires permanent at that stage. But then, you might need to do more assessment, a deeper dive, as it were, in a second round, which they call a personalized assessment where a person can produce information about how well they’ve progressed since their last conviction and then make a second decision. So, a lot of companies are trying to figure out ways to institute those kinds of policies. So, that’s a big change in how companies are approaching issues.
I’d say one last thing, is that a lot of the attitudes about hiring people with records depends critically on the tightness of the labor market. So, it’s been well established that basically if you queue up people with records and people that are on welfare and people that have long term unemployment, a number of people with different issues in the labor market, the people with records are at the back of the queue.
Harry Holzer is the leading researcher in that space and he’s demonstrated that very clearly and everybody that studies that finds that. But the key is when you have a tight labor market like we do currently, you get to the back of the queue. So, we did research in 2017-18 time period data from that period and showed that 46% of the unemployed men, in this case unemployed, means actively looking for work, had an adult conviction. So, that’s not going to be true during the Great Recession. During a Great Recession it’s going to be much lower, but during a tight labor market, during that time period, the people that are left are the people with records. So, employers start to try to figure out ways to hire them because they need workers. That puts us in the situation we are currently where if we can figure out how to hire more people with records, you get a win-win-win, which is a phrase used on this podcast before, which is employers get workers that they need, people who have records get jobs they want, because a significant proportion of them are looking for work. We as a society benefit because there’s been evidence to show that people that are actively looking for work with records, if they are hired and then fired because of the Ban the Box situation, they will then be much more likely to reoffend.
Matt: Absolutely. Yeah, that makes perfect sense. Let’s dive into it a little bit deeper because there are some real complications around this in terms of a lot of misconceptions that people have around background checks and hiring people with criminal backgrounds. Talk us through some of the common misconceptions you see amongst employers when they’re looking into this area.
Shawn: So, there’s two big ones, and I wouldn’t say it’s just employers. It’s pretty much everybody. The first one is, there’s a common set of statistics that come out of the United States, but it’s not just the US, that say that most people with a record who have a conviction or have been released from prison are going to reoffend. So, oftentimes in this space, you’ll hear, “Well, 60% of the people are going to get reconvicted, they better get a job, or they’re going to continue active in crime.” That’s actually not right. The reason is, it’s a little subtle, but it’s also really important is that when we’re having this conversation, the thing that employers want to know is, if I go out and hire someone who has a record, what is the probability that they’re going to reoffend while they’re working for me?
So, the sample that they’re interested in is everyone with a record. Well, the statistics that we use in this context to talk about it aren’t everybody with a record. They’re everybody that’s been convicted last year or everybody that’s been released from prison last year. Now, think about why that may not be- why that isn’t, I should say, everybody who has a record. Think about a 10-year time period. So, there are going to be some people that get convicted 10 times during that 10-year time period. There are other people that are going to be convicted once. If I only grab one year of conviction records, what am I going to find? Well, I’m much more likely to get the people that offend a lot than the people that only offended once. In other words, the samples that we use to talk about this are filled of people that reoffend a lot.
In other words, frequent flyers or people that are frequent offenders, which is a small subset of the population. If you want to get the right statistic for this particular problem, you want to think about everyone who’s ever had a conviction. So, give you an example of why these matters. We did this in North Carolina and looked at a group of people that were convicted during one year. We found that 70% of them are going to be reconvicted in the next, say, 10 years. If you look at everybody that’s ever been convicted and then say, “Well, how many times have they been convicted in the next 10 years?” The answer is 30%. In other words, only 30% of the people that’s ever been convicted will have another conviction in 10 years. The reason is that because some people get convicted a lot. That’s a fact that you can keep in your mind, that 5% of all people who get arrested contribute 50% of all arrests. That’s been known since the 60s.
So, there’s a few people that do a lot, well, those aren’t the people that are applying for jobs and trying to get work. The people that are applying for jobs, trying to get work are much more likely to be the people who haven’t been convicted much. So, the actual recidivism rates are much lower for the population of people that employers want to hire. As a result, there’s a lot of people out there that could be hired who aren’t being hired, who would be perfectly good risk. In other words, they’re not going to be any higher risk than the people without a record and they’re going to make good employees. So, we’re in a current situation where even if we don’t make more people not offend, we could just work with the people that are out there, that would be good risk if we could only identify who they are. The starting point to that is don’t overestimate their level of risk because the actual level of risk posed by most people is quite low.
Matt: I think that’s interesting. It’s such a classic case where statistics and data can be so misleading. You mentioned there was another big misconception. Talk us through that.
Shawn: Another way to think about the misleadingness of that statistic is to remember the following thing. It’s an easy one to keep in your mind. I estimate that 30% of American men, at least one adult conviction, forget about what they did as kids, by the time they’re 35. Now, do you think honestly all of those people are active offenders? Because the crime rate would be something like 20 times higher than it is if it’s true. So, crime is essentially young man’s game. Most people who offend do it in their young early 20s, and then they stop. So, the reality is the vast majority of people stop. You kind of know this once you realize that 30% of American men were going to have a conviction record. So, the goal is to identify the people who have stopped. Now, how do you do that?
Well, the thing that people do a lot is they focus on the crime type. They say, “Oh okay, I have a job as a cashier. Well, I shouldn’t hire anybody that’s been involved in money crimes. Okay. But maybe if it’s violent, it’s okay.” People typically don’t say that, but let’s just say, the idea being that they look for an alignment between the crime type and the job. In fact, those are the two of the three things that the EEOC said you could look at for years and years and years, they called the green factors. They said you could look at type of crime, type of job, and time since last conviction. Well, the problem here is that most people who commit crime don’t specialize, they’re generalists. So, knowing something about what they did before doesn’t really help you tell what they’re going to do next. There are two big things that drive predictions in any risk tool used in the criminal justice system, it’s age and the number of prior convictions.
Once you control for age and the number of prior convictions, type of crime actually doesn’t predict anything, people don’t specialize. So, one of the big problems is we have a system driven by type of crime, yet it isn’t actually predicting the thing that the government says you are allowed to try to avoid. So, if you want to predict risk, you want to focus on age at the time of the last conviction, which is allowed the number of priors and the time since last conviction or released from prison, not type of crime. Now, a reasonable question, Matt, is to say, “Well, why do we think it’s type of crime?” It’s a well-established psychological phenomenon called hindsight bias. You say, “Oh, okay, suppose that I walked on to the job and I punched someone in the head and knocked them out. You then looked at my record and said, ‘Oh, he had a violent crime, I should have known that he was going to be violent.’”
Well, the problem with that is just go back to the statistic I started with, which is there are a lot of people that committed crimes that never commit another crime. Well, you don’t get to see those people. In other words, you only get to see the people that had one before. It turns out but when you’re trying to do prediction, you got to look at everybody that ever had a violent crime and then ask how many of them had another one and the answer is not very many. So, hindsight bias is a source of bias where you overestimate the predictability of an event, and it’s a real problem that leads to things like negligent hiring lawsuits and other things. And it’s a real phenomenon. So, one of the things I talk to about companies is, yeah, it’s true that you can’t use science to use type of crime, but you might end up using type of crime anyways in some way because you know it’s a fact that leads to problems with negligent hiring lawsuits and it’s also an issue for brand awareness.
The average citizen is going to say, “Well, you should have known that this guy had a sex offense and was going to do a bad thing on the job if he allowed it in a retail store with waiting rooms and dressing rooms,” even if that’s actually not true.
[ad]
Matt: So, what advice do you give to employers? How can they make a factually informed decision around all of this in terms of future behavior and making the right decisions about hiring?
Shawn: If we go back to how I talked about the new framework for background checks. So, there’s this first decision that gets made where the company will do a background check and find out that say 30% to 40% of the people that they made an offer to will have a record. The goal there is to identify people for whom you don’t have to do any further background checking, you can just hire them. In other words, you don’t have to ask them about what they’ve done since last, you don’t have to do anything. You pretty much know that they’re low risk. That situation is different than what we’re used to thinking about in a prediction exercise because we’re actually predicting low offending, not high offending. It’s actually really hard to predict high risk offenders. It’s a rare group. So as a result, the only way you can justify using a prediction model in that case is if you’re willing tolerate lots of false positives i.e., people that would make perfectly good employees that you don’t hire in order to avoid the one false negative, which is the person that would make a bad employee that you then go on to hire. That’s where we have a problem as a society.
We don’t want 1000 false positives, we want those folks to work. So, instead of focusing on trying to predict the high-risk people, what I say you should do in that first step is predict the low-risk people of which the majority of people in the pool are going to be low risk. That’s actually something you can do pretty easily. It’s easy to predict more common things. That’s just intuitive. If the question is, “Are you going to wake up tomorrow morning?” I’m pretty confident I can predict that for you just by saying, “Everyone’s going to wake up and there’s a few people that won’t.”
So, the more common it is, the easier it is to get it right. Around a 50%-60% margin, what we’re talking about is a really good place not to make a lot of mistakes. So, I think that using information like time since last, number of priors and the age at the time of last conviction really allows companies to say, “Look, I don’t need to look further. You’re good. You’re no more risk than a non-offender, particularly a non-offender that I would otherwise hire,” say, a young person or something. So, that’s the first step. The second thing is, when you get to the second stage and you see someone who just came out of prison, you can actually believe your own eyes. In other words, the question will be, “Did they stop? Well, are they the same person that committed the crime in the first place?”
So, if you see that a person has graduated from a number of programs, has worked for a year or two successfully, you don’t have to worry about that person being the same person they were before, because they’re not the same person, that what there was before, that person didn’t go through programs, that person didn’t have a job. All too often you see companies saying, “Well, this person has a 10-year-old record. It was a very serious crime. Yeah, I know. I hired them three years ago and they were great employees for three years, but I’m not going to promote them because they–” believe your own eyes, you know a lot more about this person than you did before, and that person is no longer the same person. And that’s not surprising because people change. Remember, lots of people get involved in crime when they’re young and then they stop.
So, I think there’s this lack of acceptance that you can learn new things about a person. In the context of criminal justice, that’s called dynamic risk assessment. But that information about what they’re doing now tells you a lot about their current levels of risk and it’s far more informative than what they did 10 years ago.
Matt: Absolutely. I always like to finish with a question about the future. In this context, there’s obviously a huge amount of misconception out there and also missed opportunity for people looking for jobs and for employers and for society as a whole. So, how do we drive change at scale here? How do we get that win, win, win? Is there a role for government? Is it incentives? Is it better education for employers? Is it all of those? How do we drive the change?
Shawn: So, we’ve been thinking about that a lot. Some of these facts have been known for a long time, but they don’t seem to be integrated into the way we do standard practice. So, a couple of things that we’ve thought about is, the first is to recognize that there are a lot of people that are in fact what we call Desistors. They have stopped. The goal there is just identify who they are. You don’t need to make them desist. You don’t need to do anything except figure out who they are. There we can use better information, better practice in terms of actually translating the information you get into a decision, so that’s the decision rules used by the companies and things like that. And not rely so much on type of crime or other factors that don’t matter.
The second thing is there are going to be some people that are on the margin here, and the win, win, win, remember, is that if we hire people, there’s going to be less crime. But the employer doesn’t capture that. That’s why it’s called a positive externality. Meaning that the employer doesn’t get all of the benefits. Society gets some of it. Well, if society gets some of it, we can share some of it back to employers to encourage them to hire more people, because in a situation like that, they’re going to hire fewer people than society would like them to. So, there are incentive programs that the government could run, like Work Opportunity Tax Credit in the United States or The Federal Bonding Program in the United States that are currently in existence but not used much. Some of the reason why they may not be used much is because they’re inconsistent with Ban the Box. So, we’ve created a new policy to try to hire people with records, but it is inconsistent with the incentives that we then provide employers.
So, trying to find ways to make those incentives more practical, more usable would-be part of the game for getting that win, win, win. We can get a win, win, just by making better decisions about the people who have records. Many of whom are at very low risk for actually committing another crime and would make perfectly wonderful employees.
Matt: Shawn, thank you very much for talking to me.
Shawn: Ah, you’re welcome. Thanks so much for having me.
Matt: My thanks to Shawn. You can subscribe to this podcast in Apple Podcasts on Spotify or via your podcasting app of choice. Please also follow the show on Instagram. You can find us by searching for Recruiting Future. You can search all the past episodes at recruitingfuture.com. On that site, you can also subscribe to our monthly newsletter, Recruiting Future Feast and get the inside track about everything that’s coming up on the show. Thanks very much for listening. I’ll be back next time and I hope you’ll join me.
[Recruiting Future theme]
