Measuring Agreement Between Egos and Alters

In personal network studies, respondents (egos) are asked to report information about members of their personal network (alters); egos respond based on their perceptions. Previous studies of informant accuracy present a varied picture: Some find egos' reporting on their relationships with alters to be accurate; others do not. In our study of people living with HIV/AIDS in Uganda, egos were asked to invite up to four alters named during their interview to answer questions about their relationship with ego. Using Gower's general coefficient of similarity, we calculated a measure of accuracy both by variable and by alter. Our analysis by variable tends to confirm informant accuracy research, while our analysis by alter adds to the literature by identifying how accurate a particular ego is when discussing an alter and what characteristics might be associated with accuracy or inaccuracy.