This study explored the relationship between level of expressed emotion (EE) and level of family burden in 20 families of individuals with psychotic disorder. Results indicated that a relationship between EE and family burden did exist, but it appeared to be strongly influenced by demographic and illness-related variables. The usefulness of the 5-min speech sample method of assessing EE was limited by the aversion of some family members to being videotaped and by denial of patient symptoms by some relatives. Other methodological weaknesses are discussed. The relationship between demographic and illness-related variables and level of EE needs further exploration.