Measurement issues in family assessment: the case of the Family Environment Scale.

The internal consistency reliabilities of 5 subscales of the Family Environment Scale were re-examined in a study of 311 stressed and 74 control families. The reliability coefficients generated were lower than those originally reported for this instrument; most coefficients generated were below the acceptable level for practical or research use, and there was considerable variation in the reliabilities across type of stressed family. An attempt to generate more reliable scales using the original items was unsuccessful, and questions about the validity of the subscales were raised. Our results illustrate the importance of examining reliabilities of instruments, even well-known and widely used instruments, for each sample studied.