Use your words: Written prisoner interactions predict whether they’ll clean up their acts -- ScienceDailyThe first, called "pushups," are congratulatory notes to a peer -- something like, "Good job talking about your triggers in group today, man." The second, called "pull-ups," are meant to steer a fellow prisoner toward better choices -- something like, "Hey brother, next time try talking to me instead of getting into a fight." Once approved as appropriate for group consumption, the written notes are typically read aloud to the group during meal time or a meeting. Doogan and Warren examined how these communications changed for each of 2,342 men included in their study. They looked at pushups and pull-ups in each inmate's first two to three months in the program and held those up against the messages they sent fellow prisoners in the second two to three months. In all, the researchers analyzed about 267,000 messages. Only graduates of the program were included in the study. The more their word combinations shifted, the greater the chance the men didn't return to prison. In cases where the inmates did return, those who showed the least change in how they thought and wrote tended to return to prison most quickly. The study didn't focus on "positive" or "negative" word choice, but on change in general, with the goal of getting a handle on whether the program was reshaping the participant's way of thinking, Doogan said. "It wasn't so much sentiment, but whether we could measure some form of change in the individual," he said. The sheer number of interactions for an individual resident didn't seem to make a difference -- only the changing nature of those notes. That's important because it seems to mean that simply interacting isn't enough and that a person has to be engaged and evolve in his thinking, the researchers said. Shifts in how we put together our thoughts and express them in writing are a good indication of a true evolution in how we think, Warren said. "Learning is a change in connections between ideas," he said. "In a therapeutic community, you would hope that they are abandoning some old connections and developing some new ones." The researchers created a tool for analyzing word choices, identifying 500 words that could potentially be combined in a note to one participant from another. Doogan and Warren counted change when inmates added new word combinations or abandoned old ones. They attempted to control for variables outside of changed language including race, age and education level. Understanding -- and being able to measure -- changes linked to reduced rates of repeat incarceration could eventually help program directors refine how they approach different participants, the researchers said. For instance, if it was clear an addict's communications with others in the program were not changing in nature, it might be a clue that the individual needed more one-on-one attention, Doogan said.