Encountering Correctional Populations. Kathleen A. Fox

conducting program evaluations may also need to convince correctional leaders and staff to use random assignment, which means to randomly assign offenders to experimental or control groups. Using random assignment means that subjects have an equal chance of being assigned to either group. Random assignment can occur by doing something simple, such as flipping a coin, or by using random number generators or a computer program with this capability.

      Random assignment is considered the “gold standard” for evaluation research (Boruch 1997; Farrington 2003), but it can be difficult to get agencies or their staff to understand the need or to understand that random assignment is legal and ethical. That is, staff may want to use assessments or staff knowledge of client needs to determine who gets access to special programming. In our experience, for example, SOCP staff often wanted to use their personal knowledge about a youth or experience over time in the field to help determine which youths were assigned to which programming rather than rely on luck of the draw. That is, staff may be personally convinced that particular youths need a new or special program and want to bypass any procedure that would leave that assignment to chance. We have found that two points can be convincing to staff. First, random assignment is the fairest way to put youths into groups, if the program cannot serve all people who need the help it provides. It prevents favoritism and ensures all have an equal chance of being selected. Second, random assignment allows for the strongest results, meaning that agency leaders can put more faith in findings than if it were not in place. They can better sell the results, whatever they are, when asking for more funding, for example. Additionally, in some cases, if we had enough degrees of freedom, we allowed a small number of youths to be sent to the project each year (e.g., fewer than 10 percent) while bypassing the random assignment process. These youths, then, are just not to be analyzed for the study but are served by the program.

      In our experiences collaborating with staff, we found that some were easier to work with than others, which we expect is typical of most correctional studies. It was important to get to know the personalities and preferences of both the facility and agency administrators and their staffs, because it helped ease the working relationship as we periodically arrived to conduct research. The key, though, was working within the constraints of the facility or agency while trying to maximize our ability to conduct rigorous research. In other words, we were very willing to compromise on details (e.g., times, dates, places to collect data, and managing other idiosyncrasies) of each facility and its staff as long as it did not jeopardize methodological rigor.

      With some staff members and in some situations we could be more open about ourselves (e.g., our religious preferences or personal lives) and with others we felt we needed to be more guarded to encourage a smoother working relationship. In our Florida Faith and Community-Based Delinquency Treatment Initiative project, for example, religious beliefs were part of regular conversation among staff, because faith was such an integral part of the program. This meant we had to be especially attuned to our attitudes and how we expressed them, and we could not show judgment when they expressed religious or even political attitudes with which we disagreed.

      In this faith project, we also had to be careful about how we responded to behaviors we personally thought were not appropriate. For example, in one facility, we regularly saw areas in physical disarray, such as unmade beds and clothes strewn around, youth jumping across furniture, and expired food (which together signaled to us a lack of good management), while in others important paperwork was not appropriately stored. In one facility, home addresses were listed on youths’ bedroom doors, which we thought might be a safety issue for the residents. This facility was later shut down for many problems. While we jotted issues such as these in our research notes, we did not express judgment or report these sorts of problems to their agency managers because we wanted to maintain the working relationships and understand the facilities as they were, not just as they were supposed to be. We allowed the quality assurance process within the agency more broadly to manage the agency’s rule enforcement for these issues. Of course, had we seen something abusive or illegal, we would have reported it to the appropriate authorities. If the research team sees something concerning, they should be encouraged to report it to the lead researcher or principal investigator, who may wish to call the university attorneys to inquire about the appropriate course of action (if any) for the researcher.

      CONVINCING THE TARGET POPULATION TO PARTICIPATE

      Once researchers have obtained agency permission to collect data and secured buy-in from the staff, the next challenge researchers face is often how to convince people to participate. The following is a list of specific tips that have worked for us to build and maintain rapport with participants:

      •Make eye contact, be friendly, and treat participants with respect.

      •Be careful to use nonjudgmental language, including body language (e.g., avoid crossing/folding arms across your body or making faces).

      •Make it clear that you are there because you want to learn from them, that you know they have things they can teach you.

      •Reassure them about the informed consent procedures regarding your research and remind them that findings will be reported in a way that does not identify them personally or get either officers or offenders in trouble, unless they threaten to hurt themselves or others.

      •Create a comfortable research environment for participants (e.g., minimize distractions, ensure their privacy). Any research situations in which correctional officers appear to punish or negatively regard participants must be immediately terminated to protect participants. Then, the incident will need to be reported to the IRB as an adverse event.