Joe Fuller: Many employers rely on outdated hiring practices based on invalid assumptions. High on the list is the expectation that a qualified job candidate, by definition, has a history of full-time employment in their chosen field. The picture, of course, is far more varied and complicated than that, including workers with episodes of part-time and freelance work, career breaks, and career switches. This “W-2 bias” also excludes marginalized groups. What’s behind the insistence that job candidates fit in such inflexible templates?

Welcome to the Managing the Future of Work podcast from Harvard Business School. I’m your host, Harvard Business School professor and visiting fellow at the American Enterprise Institute, Joe Fuller. My guest today is David Pedulla, professor of sociology on the faculty of arts and sciences at Harvard University. David’s research delves into how nonstandard modes of work affect individual’s employment prospects. He also examines how labor markets produce different outcomes by race and gender. His book, Making the Cut, provides an inside look at how employer’s filter job applicants by employment history, gender, and race. Through his innovative field work, David sheds light on how the inferences gatekeepers draw from nonstandard resumes penalize qualified candidates. What can business leaders and policymakers do to break this counter-productive cycle? Welcome to our podcast, David, and to the Harvard faculty as well.

David Pedulla: Thank you so much for having me. I’m really looking forward to the conversation today.

Fuller: David, when I first started looking at issues of the workforce, skills gaps, diversity, I did not really understand the extent to which the study of it is spread over multiple disciplines. But now that I’ve begun to approach my second decade in studying it, I’m getting a better appreciation for all the different topics that abut the issue and influence our understanding of how labor markets work and how that reflects itself in individual opportunity. You’re a sociologist, happily recently joined the Harvard faculty. Can you tell us a little bit about your journey, and as a sociologist, what attracted you to the field?

Pedulla: I was a history major as an undergraduate. After I had finished my undergraduate education, I spent a few years working in the research and policy world. So I spent a few years at the Brennan Center for Justice, at NYU School of Law, I spent some time as a New York City urban fellow in the mayor’s office. These experiences got me really interested in thinking about issues of poverty, social exclusion, economic inequality, and repeatedly brought me back to thinking of issues of labor markets, and how labor markets work, how organizations function. I was really drawn to sociology for multiple reasons. I think sociologists are great at thinking about different methodological approaches for getting traction on key social issues. I actually thought coming into graduate school that I was going to be an ethnographer and use qualitative methods to study, mainly, I was interested in issues of the suburbanization of poverty and how it intersected with labor markets. Then when I got to graduate school, I met Devah Pager, who was an incredible sociologist and was using field experiments to study racial discrimination and discrimination against people with criminal convictions. Through our conversations, and getting to know her better, I got really excited about using field experimental techniques to study key questions about labor market inclusion and labor market exclusion.

Fuller: So for our listeners who have a knowledge of sociology, which would be they had a friend in college who was a sociology major, what is it that the sociologist brings to the study of this? In reading all your work, it’s been very interesting to see how, through your lens, you are very much in the direction of the findings we have in the Managing the Future of Work Project at HBS and of some of the applied economists studying the field.

Pedulla: I think sociologists who are interested in work bring to the forefront issues of power, inequality, and how key social dimensions are different—so processes of social stratification that would be around race, gender, age, sexual orientation. As sociologists, we care a lot about how those social categories play out in different domains. So that’s one of the things that I think sociologists bring to the study of work. And labor markets is really an emphasis on the differentials in power and the differentials in status between groups in the United States. A couple other areas where sociologists are really keen is to think about issues of social networks and the connections between people, and how those influence, say, hiring or job search or other processes in the labor market. Then lastly, what I mentioned a little bit earlier is that sociologists have a really broad methodological toolkit. So oftentimes a sociologist, we try to think about what type of method is going to provide the best type of data to answer the question that we’re interested in. So in some cases, that may be ethnography or participant observation. In other cases, that might be in-depth interviews. Or in the case of some of my work, that may be using field experiments to get at the causal effect of some category on a particular outcome of interest.

Fuller: One thing I noticed in your historical research that was quite interesting was that the impact that individual’s social networks can have on their ability to find and keep work is quite uneven, and it actually stratifies by race a bit.

Pedulla: Absolutely. So I have a paper that came out in the American Sociological Review in 2019 that was looking at how individuals use their social networks during the job-search process. In this collaborative work with Devah Pager, who was my mentor at Princeton who I mentioned, what we find is that Black and white job seekers are actually equally likely to use their networks to try to find jobs. But what we find is that the benefits that accrue through those networks are much stronger for whites than they are for African Americans. White job seekers are much more likely to know somebody at the companies they’re applying to. Also, white job seekers are more likely to have their network alt or their network tie mobilize resources or put in a good word for them than are African American job seekers.

Fuller: Yes. We’ve been very interested to look at some of the social entrepreneurs that have taken on that challenge, particularly in high techs, like Co-op Careers and Code Path, that are working backwards from that network effect to say, “We have to create a cadre of African Americans in tech, and specifically in computer science, so they can benefit from those social networks.” You published a Harvard Business Review article recently focused on strategies companies can and should follow to enhance their diversity and inclusion. This is, of course, a topic that is just absolutely in the forefront of boards of directors’ and C-suite executives’ minds. I think there is a new level of commitment, an openness to inquiry and experimentation by companies fomented by the annus horribilis of 2020. Talk a little bit about your recommendations to them, and how companies should be approaching this. What has to change?

Pedulla: So back in 2018, again with Devah Pager, we were thinking about these issues of diversity, bias, discrimination. We held a convening here at the Radcliffe Institute, where we brought together a bunch of scholars with the key goal of thinking about solutions to bias discrimination and strategies for increasing diversity, equity, and inclusion. Really focused on the solution side. Out of that came a report that I edited, and where we had a bunch of the people that were at the conference and some others come together and write different sections about these key recommendations. So the HBR article that I wrote tries to distill that longer report into a few key takeaways. So a few that I’ll highlight are, one of the first chapters in the report is thinking about the key role that data can play. The recommendation here is that companies should set clear goals and benchmarks that they want to hit with data and then collect the right metrics over time that are measuring how they’re performing on those different key outcomes. But it’s really important to share out that data with key stakeholders. So key individuals—internal and external to the organization—can be made aware of those benchmarks, can see the data, can cut the data in different ways. That serves as a really important way for organizations to increase diversity, equity, and inclusion, and in a lot of ways to hold themselves accountable to the goals that they set out. Another key area that the report explores was thinking about issues of technology used in the hiring process to sort and screen applicants using algorithms and other sorts of automated processes. So one of the key aspects of this report was to proactively think about how to create and utilize technologies in ways that are fair. Tests around algorithms and fairness need to be from the beginning built into the design of the algorithm and then continually updated and monitored. Then the last point is the importance of organizations bringing in managers right from the start. So oftentimes at organizations, there is this idea for increased diversity, equity, and inclusion. You’re kind of layering additional work onto what managers are already doing. It’s not necessarily grafted into how a manager would have designed it to begin with or what they’re dealing with on a day-to-day basis.

Fuller: A couple of things in the article that were very resonant with me and consistent with our research. One is to have very visible and committed senior sponsorship, and it’s very hard to start your way down the path to better levels of diversity and inclusion if you don’t have any leaders that are diverse, who are role models, and can both help the company learn what it’s going to take to be more successful in their DEI efforts and also be points of connection for those younger workers to create a vibrant pipeline of diverse talent that will then become a self-sustaining credit to the organization. Let’s turn our attention to your book, Making the Cut, because that also speaks to issues of diversity, inclusion, and to me, very interesting in that you engaged hiring managers to try to understand how they were processing information, and what kind of judgments, implicit or explicit, they were making in evaluating candidates. Talk to us about the book, and your findings.

Pedulla: Early on in my academic career, I got really interested in thinking about employment relations in the United States. So we often think of the kind of old school, standard, ideal typical employment relation of a full-time job. We know that that type of employment relationship was not the case for many workers in the United States ever and is not the experience of many workers today as well. So I got really interested in thinking about things like part-time work, temporary agency employment, what I refer to as “skills underutilization”—or people in jobs below their skill level. You can think about other types of nonstandard employments, things like gig jobs, independent contracting, etc. I began to realize that the data in the United States to study these sorts of things is not great. We don’t have really good data on things like temporary agency employment. So I decided maybe I would collect my own data. In particular, the key question I was interested in was, given that so many workers have these different types of nonstandard or precarious employment histories, how do employers evaluate them? If we look at the demand side of the process—where employers are evaluating job applicants who have these employment histories—how do they make sense of them? I had drawn two main types of data: One is using field experimental techniques, where I send out fictitious job applications to apply for real job openings and randomly assign different applicants to have, say, a part-time history or a temporary employment history. I also overlaid the race and gender of the job applicant using names on those different resumes. I had a real interest in these field experiments where you can identify the causal effect of one type of employment experience on getting a callback for a job in the actual labor market. But then, with those types of techniques, you can get these nice, quantitative causal estimates, but you don’t necessarily know what are employers thinking. How are they making sense of these employment histories? I said, you know what? I think I really want to talk to hiring managers, to recruiters, to people who are involved in the staffing process and see how they make sense of these different types of employment histories. I interviewed over 50 hiring managers and recruiters. We conducted in-depth interviews. I worked with two research assistants at Stanford on this project. It was incredibly revealing to get a sense of, how do these hiring managers actually think about different types of employment histories? What matters to them? In addition to them ascribing meanings to these employment histories around someone’s personality or their soft skills or whether they’re committed or competent, those types of things, one of the key findings was that having a nonstandard employment history in general induces a huge amount of uncertainty. So when you’re reviewing a job application, and you see a part-time history or you see a spell of unemployment or you see someone working a job below their skill level, hiring managers really wanted to know, “Why is this the case? What’s going on?” That makes sense, right? It was really helpful to talk to the hiring managers about how they would then, given the uncertainty or given the questions that arose, then how they would decipher whether this was someone to call back for a job or not.

Fuller: It’s fascinating, as you call out in the book, how quickly hiring managers draw inferences from job histories and, through that cascade of inference, make judgments. We have certainly seen in our research how those types of inferences get committed to everything from job descriptions to the filters and screens in applicant tracking systems, and then create what amounts to what the great Chris Argyris callrf “closed-loop learning.” You’ve built an assumption into your system, you’ve put that in the system. The fact that it’s there is undetected, and the fact that it’ll lead to biased outcomes is undiscussable. It’s a fascinating loop. When you were talking to these hiring managers and you looked at the data, what did you conclude about the balance between people are taking shortcuts and trying to be efficient and that has this consequence, versus people are suffering from subconscious bias and that causes them to make these judgments?

Pedulla: So the power of the types of field experimental techniques that I used in the book is that I was able to hold everything constant about the job applicants. So they had the same types of educational credentials, similar types of employment trajectories, until the most recent employment history is where I randomly assign some workers, say, a part-time job versus a full-time job. So I can see the kind of unique inferences that employers are making based off of just that one characteristic or that one category, and then can further see how that varies with the race and gender of the applicant. So what I’m able to detect is, really, what are these subtle processes that are leading to the types of evaluations that I pick up in the data?

Fuller: It’s going to be very interesting to see how this evolves in the minds of hiring managers, because certainly our research published late last year indicates real growth in the types of labor platforms that highly skilled workers can employ to ... It’s going to be very interesting, also, how this bias against nonstandard working relationships play out, because one of the most dynamic sectors in the gig economy is highly skilled, highly qualified workers, worked on platforms like Catalant and A-Team and Toptal and Braintrust, who have exactly the type of skills that are in severe short supply, and companies are dying to secure, but where the workers are finding both the total compensation, but also the quality of work life, superior. So do you imagine that we’ll have an evolution in this bias or that it’ll start to stratify, discriminating against lower-skilled, lower-educational attainment workers on those type of gig platforms, versus our PhD computer scientists on the A-Team?

Pedulla: There have actually been some more recent audit studies and field experiments looking at these types of issues. So there was one study that recently was published in Social Forces looking at people who took time as freelancers, right? What does that signal if you’re applying back into the labor market for a job? What does that mean to employers? There is some finding of stigma and bias against people who have freelancing experience. There is another recent paper that came out that finds similar bias against entrepreneurs—so people who started their own business and then are coming back and applying for jobs. They actually find, I believe, that those negative effects are stronger for men than they are for women, which his interesting, right? Bringing these social aspects to those penalties. So I do think there is a lot of interest in tracking and monitoring these processes over time, thinking about how they might change. One of the things that I think is going to be fascinating is, as particular types of employment become more common, and maybe people are transitioning back and forth between these highly skilled platforms and organizational employment, we may see changes in the types of stereotypes or associations that hiring managers and recruiters have about these types of positions. I think that’s going to be a really important set of empirical questions to continue to monitor, to continue to track. Maybe high-skilled independent contracting is not going to be particularly penalizing for some groups and quite penalizing for others. This is where I think, as sociologists, there is a lot of interest in making sure that as we run these studies, and as we see how things unfold, that we keep those issues of race and gender and sexual orientation, immigration status, etc., at the forefront.

Fuller: It’s interesting, also, in your research how you highlight that some of the inferences that hiring managers use to make judgments can be applied to people who otherwise are highly credentialed—like, as you say, someone who has been an entrepreneur but may have a college degree or even a graduate degree, where historically that type of credential has been a huge boon to people in terms of getting employment. What we’re seeing, as we both survey and talk to companies, is that the velocity of change and work processes in technology now is so fast that they do begin to assume that if someone hasn’t been at the coalface of whatever industry they’re in the last year or two, that they’re going to be behind. That creates a very strange, I think, troubling challenge—which is, increasingly, to have the skills to be considered by a leading company for a state-of-the-art role, you’re going to have to be employed in that role or some feeder role—coming either back into an industry or getting an entry-level job at an industry where you don’t have that regular exposure to work practices and processes and technologies of today is going to be increasingly difficult.

Pedulla: I think that’s exactly right. As workers are moving between organizations, moving between jobs, I think employer’s expectations of what a new employee can bring to the table are kind of ratcheting up and ratcheting up. I think we really need to think about what are the consequences of that going to be for workers moving into new positions, for people who need to take time out of the labor market for a host of reasons—whether they be caretaking, parenthood, caring for an elderly parent. As we think about the various responsibilities individuals have outside of the workplace, it does raise some concerns for me that, if you have to take a short break from work to deal with something else, that that’s going to be a real increasingly difficult thing to overcome for workers who are looking to get back in for higher-paying jobs.

Fuller: I wanted to talk about your definition of suboptimal work, and how Covid affects that in a couple of ways. You describe suboptimal, as I would put it, through the eyes of the employer. This person seems to be taking the role, as you just illustrated, or took a demotion or jumped ship into outside the industry to some role that I don’t put much credit in and don’t feel, as a hiring manager, is terribly relevant. That’s one form of suboptimal. Now you have a lot of talented workers saying, “I have concluded that my quality of work life prior to Covid was suboptimal.” We can see that in the very reluctant uptake rates of employees going back to employers who are saying, “We want you back on prem, we want you within the four walls of our company.” We have a real spike in retirements, we have a real spike in turnover. As you think about that through the lens of your training and your research, what are your hypotheses about how that gets all sorted if these two dueling definitions of suboptimality collide?

Pedulla: I think as a sociologist who studies work and employment and social exclusion, my focus has been much more on thinking about the costs that come with workers who have been displaced because of the pandemic and recession, workers who have been pushed into jobs well below their skill level, workers who have not been able to find full-time employment and are in part-time jobs. I think your point is really well taken, that we need to think a little bit more broadly about the costs of that. There is certainly the financial and economic costs that really matter in terms of the paychecks and benefits that people are bringing home. But I think it’s also really important that we center things like social stigma, right? So what types of social experiences and stigma might people face within their social networks, within their neighborhoods, among their families for not being able to have the standard type of employment that they’ve been expected to have for so long? I think there is some really interesting research about the consequences that nonstandard employment and unemployment can have within household and family dynamics. So there is some great work by my colleague Sasha Killewald in the sociology department at Harvard that shows that, when men have anything other than full-time employment, it increases the risk for divorce, right? So we can see these ways that these social exclusion in the workplace can also lead to challenges within people’s homes and families. We know from a long line of research that long-term unemployment can have really severe health consequences for individuals, both in terms of mental health, as well as physical health and wellbeing. Then another aspect of suboptimal work is that a lot of workers, as the retail sector and other service sectors return to full strength, experience suboptimal jobs in terms of scheduling control and schedule uncertainty and the precariousness of their schedules. I think this is really important if we want to be thinking about job quality and worker wellbeing. So my colleague Danny Schneider at the Kennedy School and his coauthor and collaborator, Kristen Harknett, run the Shift Project, which is a fantastic social science project and survey project looking at a lot of these schedule unpredictability and schedule uncertainty issues for service-sector workers and how those play out. I think that’s really important to keep in mind as the recovery happens and as we work back into the workplace as well.

Fuller: So let’s shift the focus a little bit. Thinking back over your research, and thinking through the lens of trying to improve performance on diversity, equality, and inclusion, and also in the service of making employers and employment relationships more efficient and effective, what would you be saying to policymakers? Are they focusing their attention on the right questions? Are there questions that you think need to be asked and really probed at the level of the federal government?

Pedulla: So I think there are a few things that are really important for policymakers to be thinking about. The first is, I think, being really clear that we need to kind of enforce laws that are on the book and ensure that the types of protections that are out there for workers are being enforced by the relevant agencies. These are everything from issues of discrimination by race and gender and national origin, all the way to issues of wage theft and health and safety issues in the workplace. So I think there is a really important role for the federal government to step into the workplace and ensure that the laws that are on the books and the regulations that are out there are being enforced. The other thing that I think is important moving forward—and this is less on the policy side, but more something that I think is important for public opinion—is that I think it’s important that us as academics ensure that our findings are getting out there, are being noted in the popular press. So one of the things that I find often when I have conversations with people is they’ll make claims like, “Racial discrimination no longer exists. That’s not a problem in the United States anymore.” I think one of the things that we have really good social science evidence on is that, yes, racial discrimination does continue to exist, and there is some great meta-analytic work that’s looked at hiring discrimination from field experiments over the last 30 years and has shown that discrimination against African Americans in hiring has actually not declined over time. I think getting those types of findings out into the world is going to be really important to shape these conversations and get a collective fact base about what’s actually happening in the world that we can craft policy from.

Fuller: How do you think companies ought to be approaching this—both in the wake of the pandemic and generally? Because, as you just said, many companies would be, at the very least, startled and disappointed, and some resentful and dismissive of the characterization that, well, racism still exists, and it’s still evident in your hiring practices and in your workforce. Do you have specific best practices that you want them to consider?

Pedulla: So one thing that I think is really important is for businesses and companies to really think about standardization in their hiring process so we know when there is ambiguity in evaluation criteria. That’s when various types of biases and stereotypes and click heuristics are likely to be used. So both to root out race and gender discrimination, but also to kind of root out biases against, say, people who have a part-time job or long term unemployment. Also within organizations, thinking about issues of promotion, wage setting, really trying to think about what are the criteria that we as a company care about? That may vary from company to company a little bit—what the key axes of evaluation look like can certainly be adapted to meet the needs of the business environment that a company is functioning in. Then secondly, I think it’s useful when we’re using technology in the screening process—so things like applicant tracking systems, algorithmic screening—there is a lot of important work, including some that’s mentioned in the HBR piece that we talked about, about really ensuring that those algorithms are fair, that they’re not introducing bias by race and gender. But we can also think about ways that algorithms could be screening out applicants who have been unemployed or applicants who have been in part-time positions. I think if companies want to really be thinking about issues of equity and inclusion for people who have lost their jobs in the pandemic or been pushed into these nonstandard positions due to the pandemic and recession, really thinking about those algorithms, those applicant screening process, and making sure that they’re not systematically excluding the workers who have ended up in these nonstandard or precarious positions through no fault of their own in the wake of massive layoffs and unemployment that we’ve seen in the United States over the past year.

Fuller: It does certainly come back to the centrality of knowing, tracking, and reporting data, because if the data is producing consistent patterns or if anomalous, minimally, executives and decisions makers need to have the discipline to ask why.

Pedulla: I think that’s exactly right. I think it actually brings together—so the point I just made ties to the data collection, monitoring key stakeholders. I think it also relates, really importantly, to the point I raised about bringing managers in from the start and bringing in people from across the organization from the start to have a collective understanding of what metrics matter, what data is going to be important, so that you’re measuring the right stuff and monitoring the right stuff. If you have buy-in from across the organization from the start, I think you’re really setting yourself up for success and setting yourself up with the right types of tools and buy-in from various stakeholders.

Fuller: Well, David, thank you so much for joining us on our podcast. This has been very enlightening. I am looking forward to keeping up with your research and trying to find some opportunities to collaborate going forward.

Pedulla: Absolutely. Thank you so much for having me on today. It’s been great to chat, and I’ve really enjoyed the conversation. Fuller: We hope you enjoy the Managing the Future of Work podcast. If you haven’t already, please subscribe and rate the show wherever you get your podcasts. You can find out more about the Managing the Future of Work Project at our website hbs.edu/managingthefutureofwork. While you’re there, sign up for our newsletter.