| Social support has been linked to improved physical and mental well-being, especially for those facing serious illness such as breast cancer. However, little progress has been made in understanding the mechanism of these effects. This dissertation tested whether women participating in online support groups for breast cancer benefited from reading more highly person-centered messages, which have been shown by researchers to be perceived as more helpful, supportive, and involved by explicitly acknowledging and elaborating on the feelings of the distressed individual. A secondary objective of this research was to determine whether a computerized linguistic analysis tool can predict hand-scored person-centeredness, a time-consuming task when completed by hand. Each objective was tested through two studies. The effects and perceptions of highly person-centered messages were tested through an initial analogue experiment with a sample of women at risk for or with a history of breast cancer. A second study utilized data which collected through a controlled experimental group that provided access to the messages read by each woman and her baseline and outcome evaluations. Results of the analogue effects study support the positive effects of highly person-centered messages on the women reading them in terms of perceived social support and perceived self-efficacy to cope with cancer. Findings from the larger, controlled study also indicate that there may be a directionally positive effect of reading highly person-centered messages, but only if these messages are directed specifically to the individual reading them. The second objective, validation of the computerized coding, was assessed first through a feasibility study with a sample of 334 messages obtained from publicly accessible support groups that would be the source of messages for the initial experiment, and second on a sample of 400 messages collected from the controlled experimental group that would be the basis for the main effects study. Results indicated that the software tool could successfully predict person-centeredness. However, an interesting finding was that the linguistic categories found to be associated with highly person-centered messages varied greatly in each setting. |
