Mental Health Services Research

, Volume 3, Issue 3, pp 141-153

First online:

Reliability and Validity of a Modified Colorado Symptom Index in a National Homeless Sample

  • Kendon J. ConradAffiliated withDepartment of Veterans Affairs, Hines VA Hospital Email author 
  • , John R. YagelkaAffiliated withCenter for Therapeutic Community Research, National Development Research Institute
  • , Michael D. MattersAffiliated withHealth Policy and Administration, School of Public Health, University of Illinois at Chicago
  • , Alexander R. RichAffiliated withDepartment of Community Mental Health, University of South Florida
  • , Valerie WilliamsAffiliated withPolicy Research Associates, Inc.
  • , Mary BuchananAffiliated withPhiladelphia Health Management Corporation

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access


This study examined the reliability and construct validity of a modified version of the Colorado Symptom Index (MCSI), a brief, self-report measure of psychological symptomatology, in a study of interventions to prevent homelessness. Eight projects in a national, cooperative study collected new data at baseline, 6, and 12 months using a set of common measures as well as site-specific instruments. The pooled sample consisted of 1,381 persons in treatment for mental illness or substance abuse (or both), of which 84% had a history of homelessness. The analyses employed classical and Rasch methods to examine the MCSI's content validity, internal consistency and item quality, test/retest reliability, dimensionality, appropriateness for the sample, construct validity, and responsiveness to change. This 14-item scale was found to be a reliable and valid measure of psychological symptoms in this sample. Its content was consistent with other symptom measures, its high internal consistency and test-retest coefficients supported its reliability, its relationships to other measures indicated that it had good construct validity, and it was responsive to change. We conclude that the MCSI performed similarly to longer, more widely used measures of psychological symptomatology and could be useful in other studies targeting homeless adults with severe mental illness, substance use disorders, or both.

psychometrics Rasch analysis psychological symptoms homeless