Scientists Who Study Criminal Behavior

Breaking Into Criminal Behavior Research Is More Bureaucracy Than Breakthrough

I spent about seven years working as a research associate with a corrections-adjacent think tank before moving into pure academic work. The job sounds exciting until you actually start doing it. Most people confuse studying criminal behavior with understanding why someone breaks the law. It is far less dramatic and far more tedious than television has trained you to expect. You deal with messy data, institutional gatekeepers, and the chronic problem that criminals do not fill out surveys accurately. The core of this work is behavioral research methodology applied to a population that actively dislikes being studied. You will work with correctional staff who have their own agendas. You will wrangle IRB approvals that move at a glacial pace. And you will spend an absurd amount of time trying to replicate findings from small samples because one good study does not equal a reliable one. The primary methods fall into roughly three buckets. Quantitative work involves large-scale datasets like the National Violent Offender Registration and Imaging System or state recidivism records. Qualitative work is mostly interviews and ethnographic observation inside facilities. Mixed methods combine both and usually produce the most defensible conclusions for policy audiences. If you prefer controlled lab conditions, you can do cognitive experiments with parolees or offenders in community settings, but recruitment gets complicated fast.

I should note that criminal profiling, which is what most people picture, is actually a tiny sliver of this field. Only a handful of professionals in the United States work full time on behavioral analysis for law enforcement, and even those people mostly consult on ongoing investigations rather than construct profiles from scratch. The FBI Behavioral Analysis Unit is real but small. Most actual research happens in universities, state departments of corrections, and nonprofit evaluation groups.

Setting Up a Legitimate Research Practice

Before you touch any data, you need institutional credentials. If you are not embedded in a university or government agency, IRB review will block you from nearly everything that matters. Private practice researchers sometimes work around this by partnering with a host institution for oversight, but that takes time and usually means sharing authorship. The equipment list is surprisingly modest compared with other science fields. You need a secure data management setup, transcription software if you are running interviews, and a statistical package like R or Stata for quantitative work. Qualitative researchers rely heavily on NVivo or MAXQDA for coding. If you plan to collect survey data from confined populations, you will also need consent forms that meet both federal guidelines and the specific policies of each facility you approach. Those policies differ from state to state and even from prison to prison. Downloadable resources exist but they are fragmented. The National Institute of Justice publishes toolkits and methodology guides that are free to access. The Research Triangle Institute has archived instrument libraries for risk assessment that researchers sometimes adapt. There is no central hub for criminal behavior research materials, so you end up collecting them from journal repositories, government databases, and professional organization pages over months of searching.

Running a Study Inside a Correctional Facility

This is where theory meets reality and reality usually wins. I designed a study once that looked fine on paper. We wanted to interview recently sentenced offenders about decision-making processes prior to their arrest. The protocol was clean, the IRB approved it, and we had letters of support from a county jail and a state prison. The first problem was scheduling. Both facilities operated on lockdown schedules that changed with zero notice. You cannot plan interview windows when the unit goes into lockdown because of a fight, a medical emergency, or a counts error. I built in forty percent extra time on my original timeline and it was still too tight. The workaround was to shift from in-person interviews to phone interviews for about half the participants after I proved to the facility administrator that we could verify identities remotely without compromising security. That opened up scheduling flexibility and cut data collection time from roughly fourteen weeks down to about six. The second problem was consent fatigue. Informed consent takes longer in correctional settings because you have to explain voluntarily participation without penalty in language that passes legal review. Some participants understood it. Many did not. A few signed forms without reading them because they assumed compliance was expected. I started recording a brief audio summary of the consent process and playing it before handing over the paperwork. That reduced miscommunication issues noticeably and gave me a cleaner audit trail if anyone questioned the process later.

The third problem was data contamination from the environment. Correctional facilities are loud, distracting, and not private. I learned this the hard way when I realized half my interview data included contextual noise that made certain questions unanswerable. The fix was moving to individual cells for interviews when possible and using portable sound-dampening panels we brought from our university. That sounds unnecessary but open hallway interviews produced unusable recordings for about twenty percent of sessions. Panels changed that to under five percent.

Pitfalls That Will Waste Your Time

Sample bias is the first major trap. People who agree to participate in criminal behavior studies are not representative of the broader population. They tend to be more educated, more literate, and more motivated by external incentives like study credit or small payments. That skews findings toward risk-averse perspectives and away from the impulsive decision-making your research might actually be trying to capture. Another issue is the replication crisis hitting this field harder than many realize. Criminology relies heavily on longitudinal data that takes decades to produce, and funding cycles do not match that timeline. You will see a lot of published work based on samples that are thirty years old because nobody funded a fresh cohort. When you design your own study, account for attrition and plan for follow-up intervals that are realistic rather than ambitious. A counter-intuitive point that beginners often miss: self-report data from offenders is unreliable, but administrative data is also unreliable. Arrest records miss unreported crime. Incarceration records miss people who never get booked. Self-reports miss shame and strategic lying. The most defensible work triangulates across sources instead of treating any single dataset as truth.

Tools and Software Worth Learning

R with the tidyverse and lme4 packages handles most quantitative work in this area. If you run multilevel models with offenders nested within facilities, R is straightforward and free. Stata is common in government-adjacent research and has excellent panel data commands. SPSS is still used but mostly in older labs that migrated late. For qualitative coding, NVivo and Dedoose are the main options. Dedoose runs in a browser and works better for distributed teams. NVivo is more feature-rich for individual researchers. ATLAS.ti is another choice but I find it slower for large interview corpora. Data visualization matters more here than in some fields because your audience includes policymakers and judges. Learn to make clean tables and straightforward graphs. ggplot2 in R handles this well once you get past the learning curve. Avoid decorative charts. Nobody reading a recidivism report cares about your color palette.

Where to Publish and Get Found

Criminology journals tend to be specialty-focused. The Journal of Quantitative Criminology favors methods-heavy work. Crime & Delinquency publishes broader studies. Justice Quarterly is another solid option for policy-relevant research. If you work in forensic psychology, Law and Human Behavior or the Journal of Forensic Sciences are more appropriate venues. Preprint servers are becoming more accepted in this space but not universally. Some funding agencies and institutions still view them negatively. Check your target audience before uploading. The Open Criminology platform is growing and designed specifically for open-access work in this area.

How Scientists Who Study Criminal Behavior Influence Policy Without Losing Credibility

The tension between research and policy is real. Policymakers want clear answers. Researchers have to say the data is noisy and conditional. I resolved this by writing two versions of every major finding: one paragraph for executive summaries that states the direction and magnitude without hedging, and one section in the discussion that lays out the limitations and competing explanations. It keeps both audiences satisfied and protects you from being cited out of context, which happens more often than you might think. Another practical tip: build relationships with state-level evaluators early. They commission studies, control access to datasets, and can make or break your career depending on whether you produce work that survives peer review while also being useful enough to fund the next round. I stopped chasing federal grants exclusively after my third failed application cycle and shifted toward state contracts. The pay is comparable, the timelines are shorter, and the data access is usually better because you are working directly with the agencies that hold it.

When This Field Is Not for You

If you expect clean datasets and quick answers, leave now. The work requires patience with bureaucracy, comfort with ambiguity, and the emotional stamina to read about violent crime repeatedly without numbing out completely. Burnout rates in this area are higher than in adjacent social science fields because participants often have trauma histories and the subject matter does not let you step away mentally. If you prefer controlled environments and predictable schedules, quantitative epidemiology or public health research might suit you better. You still deal with messy data, but the institutional barriers are lower and the timelines are tighter. Criminal behavior research rewards people who can tolerate slow progress and occasional dead ends. The field is smaller than pop culture suggests but more important than most people realize. Good work here changes sentencing policy, improves rehabilitation programs, and reduces harm. Bad work changes nothing and wastes money. Design your studies carefully, document your methods transparently, and do not overclaim what your data supports. The rest takes care of itself.