Research Corner: Outcomes of Roommates Groups
Amanda Surdock, M.A. & Shannon Hill Ph.D.

As we have reported in previous issues of the newsletter, Education and Research staffers have been studying an intervention designed to show our residents how to interact appropriately with each other and prevent some of the con flicts that invariably arise when multiple people reside together under the same roof. The groups met for six weeks and discussed topics such as being a good roommate, recognizing one's own bad habits, conflict resolution, and fighting fair. During each session, the group members would discuss the topic, watch a television or movie clip illustrating both good and bad role models of the topic, then have homework where they demonstrated the skills they learned in session.

After completing the groups, we have col lected all of our data and crunched numbers to see what we have learned from running these groups. Overall, the residents reported that they enjoyed the class, and they liked being able to discuss their problems in a group. Also, the more sessions they attended, the more they improved in sessions. The information collected from the DSP's indicated that males tended to be rated as getting along with each other better than females, and that individuals with bet ter developed verbal skills tended to have more difficulty getting along with their housemates/roommates than individuals with less developed verbal skills.

Unfortunately, we are not able to answer our most basic question: does this intervention work? In order to say "it works" we'd need to show that those who came to group and im proved in the session also got along better with their roommates/housemates. We assessed "getting along" by asking DSPs (or, for apart ment residents, case managers) to rate each par ticipant's skills in conflict resolution and to rate how well the person gets along with each of the other individuals with whom he or she resides. Since the study began last February, there have been 37 moves among the resident population (admissions, discharges, moving among hous es, and one death). Since we required staff to rate "getting along" with every individual in the house, each one of the moves complicated our data. Moreover, 14 of the residents involved in the moves were group members (roughly 20% of our sample). This makes comparing start rat ings with finish ratings sketchy at best.

Changing raters presented another complication. We were using DSPs and we asked that when possible the same person complete the ratings at all five intervals. When it comes to subjec tive issues like "getting along," ratings will vary between people, but a single person's rating over time should be more reliable. During the course of our study, several DSP changes were made. 58% of the DSPs who completed pre-test rat ings were no longer employed at post-test time. Additional staff changes were common after the first DSP left a house, as people shifted from one house to another or as new hires found the job to be a less than perfect fit.

Problems with research do have one benefit: they lead us to refine our direction and ask new questions. Change is ever-present in all our lives, but previous research has already demon strated that it is even more common for people with developmental disabilities. So, we will be looking at new ways to assess behavioral change over time that will rely less on factors that are likely to change.

Our other opportunity relates to a common problem in the U.S.; DSP turnover. At The Bad dour Center, we are blessed with a low overall turnover rate. However, DSP hiring and reten tion are obstacles we face. We are hoping that our training programs and the DSP groups discussed in the previous edition of this column will help address part of the issue. We also work closely with Programs & Residential Services and Human Resources to identify new ways of recruiting, screening, training, and retaining good staff members to work with our deserving population.