After chatting about something mundane — what do you do in a typical day, for example — the participant was asked to disclose a moderate-intensity experience that researchers had selected from the participant's list of five positive or five negative recent experiences.

"We were trying to more or less simulate the types of conversations you might have with a friend in the real world," Holt-Lunstad said.

Blood pressure rose across the board for those who brought in ambivalent friends, and not just when discussing a negative experience, although that's when it was highest.

They also measured the friend's blood pressure, but haven't analyzed that yet to see what effect the interaction has on the friend.

Holt-Lunstad didn't know who was assigned to bring in a supportive friend and who was told to bring a mixed-feeling friend. But she said she could usually tell just by watching their blood pressure and pulse.

She said they were surprised that participants who shared good news with an ambivalent friend didn't get the physical benefits that normally occur. Instead, they seemed to "disengage" from the discussion of their own good news.

Holt-Lunstad said the findings have suggested future research, including whether avoiding specific topics that are a source of conflict will reduce stress.

They also plan to study the extent to which various barriers influence maintenance of the ambivalent relationships, including social ties, family relationships or the fact that you live next door to or work with each other.

---

E-mail: lois@desnews.com