The reliability and validity of the SAS-SMI (Social Adjustment Scale for th
e Severely Mentally ill) were assessed in three samples of 670 persons with
severe mental illness from a public mental health system. The SAS-SMI reta
ined comparable factor structures (Social: Family, Work, Residential Stabil
ity, Romantic Interest, and Personal Well-being) to its predecessor, the SA
S-II, and yielded high item-to-scale internal consistency correlations, goo
d interrater reliability, and comparable convergent and divergent validity.
The SAS-SMI's performance across these dimensions makes it a potentially v
aluable tool in mental health services research investigations. Limitations
and suggestions for further development are discussed. (C) 1999 Elsevier S
cience Ltd. All rights reserved.