Test-Retest: ICC=0.86 for CIQ total, 0.88 for home integration, 0.66 for social integration and 0.80 for productivity (Cusick et al., 2000)(TBI). (Willer et al., 1993) reported r=0.93 for home integration, 0.86 for social integration, 0.83 for productivity and 0.91 for CIQ total. (Seale et al., 2002)(TBI) found r=0.63 for productive activity, 0.70 for social integration, 0.71 for home integration, 0.81 for CIQ total scores, whereas (Willer et al., 1994) found that r=0.91 for patients’ and 0.97 for family members/caregivers assessment.

Interobserver Reliability: Willer et al. (1993) reported interrater reliability between patients with TBI and their family members of r=0.81 (home integration), 0.74 (social integration), 0.96 (productivity) and 0.89 for the total CIQ score.

Internal Consistency: (Willer et al. 1993); Willer et al. (1994a) reported item-to-total correlations ranging from 0.32 (socialization) to 0.67 (housework, leisure activities). Additionally, reported values include α=0.76 for total CIQ, 0.84, 0.83, and 0.35 for home integration, social integration and productivity, respectively  (Willer et al. (1994), cited in Dijkers (1997)). Post severe TBI reported values were α=0.26 (productivity), 0.65 (social integration), 0.95 (home integration) and 0.84 (total CIQ) (Corrigan & Deming, 1995) (varying etiologies). Subtotal to total correlations were reported to be 0.54, 0.74, 0.79 for productivity, social integration and home integration, respectively (Corrigan & Deming, 1995) (varying etiologies).

Construct Validity: Three components with eigenvalues >1 were identified and maintained for orthogonal rotation. These 3 factors labeled Home Competency, Social Integration and Productive Activity, accounted for 51% of variance in the set of variables. All items loaded significantly, with the finance item moved to home competency and traveling being included in social integration, while shopping was excluded since it loaded significantly and equivalently on 2 factors (Sander et al., 1999). Dijkers (1997) reviewed 4 articles providing correlations between subscale scores and found moderate to weak correlations, suggesting that there are three dimensions which are related to each other and may be assessing different aspects of the same concept. (Kuipers et al., 2004)(ABI) reported a more stable 2 dimensional structure on multi-dimensional scaling (productivity versus personal life and independence versus dependence), although they were also able to identify a 3 dimensional structure in keeping with factors of home competency, social interactions and productive activities. Lequerica et al. (2013) compared a multicultural population with TBI and found that the factor structure of the CIQ was most suitable for the Caucasian population, less so for the Black population, and unsuitable for Hispanics.

Construct Validity (Known Groups): Willer et al. (1993) reported that a group of individuals with TBI versus a non-disabled group demonstrated significantly less integration on CIQ (total scores and all subscores) except for women who were equally integrated in the home, regardless of group membership. Differences in CIQ subscores and total CIQ scores were significant (p<0.0001) when a group of individuals with TBI and a group of non-TBI control participants were compared Willer et al. (1993). Groups of patients differentiated by independent living, supported living and institutional living setting could also be distinguished by differences in CIQ scores (p<0.001) (Willer et al., 1994)). Corrigan and Deming (1995) reported CIQ scores did not differ significantly between groups of persons with various disabilities (2 TBI samples versus “other disabilities”; p>0.01).

Concurrent Validity: Total CIQ scores are correlated with total Craig Handicap Assessment and Reporting Technique (CHART) scores (r=0.62, p<0.05) and 2 CHART subscales appear comparable to CIQ subscales (occupation & social integration) (Willer et al., 1993)). CHART occupation is correlated with all CIQ subscales and most strongly with CIQ productivity (r=0.55), while CHART social integration is correlated with CIQ (r=0.35), but the correlation didn’t reach significance (p>0.05; (Willer et al., 1993)). CIQ subscale and overall scores correlated significantly and in the expected direction with Disability Rating Scale (DRS) items and FIM+FAM items. DRS level of functioning scores correlated most strongly with home competency (-0.46) and total CIQ scores (-0.47), while DRS employability correlates with CIQ productive activity (-0.58) and CIQ total scores (-0.58). FAM community access correlates with home competency (0.46) and CIQ total (0.47), while FIM social interaction correlates with all CIQ subscales (0.24-0.27) and CIQ total (0.34). FAM employability correlates with CIQ productive activity (0.57) and CIQ total (0.60) (Sander et al., 1999). CIQ total scores correlated significantly with DRS total scores (r=-0.43, p<0.01); CIQ home integration correlated with DRS cognitive ability, level of function and employability subscales. On the other hand, CIQ social interaction and productivity scales did not correlate significantly with any of the DRS subscales. CIQ total correlated significantly with CHART totals (r=0.68, p<0.01), CHART physical correlated significantly with CIQ home integration (r=0.53, p<0.01) and social integration (r=0.25; p<0.05). CHART social interaction correlated with CIQ social integration (r=0.38; p<0.01), CHART motor correlated significantly with all CIQ subscales (r=0.40-0.47, p<0.01), as did CHART occupation subscale (r=0.33-0.42, p<0.01) (Zhang et al. 2002). CIQ subscores correlated with ratings of Activities of Daily Living (ADL) (r=0.37, 0.37 and 0.40 for home integration, social integration and productivity, respectively (Heinemann & Whiteneck, 1995) (TBI).

Predictive Validity: Heinemann and Whiteneck (1995) reported that Social Integration and Productivity subscale scores were the two most powerful predictors of life satisfaction on multiple regression (β=-0.25 and –0.22, respectively) such that greater satisfaction was associated with less social and productive handicap.

Revised Subscale & Scoring: Sander et al. (1999) repeated factor analysis resulting in a slightly modified subscale structure. Recommendations for a revised scale and scoring are provided. Using the revised scoring proposed by  Sander et al. (1999), CIQ total scores were significantly related to CIPI social activity and inactivity subscales (r=-0.43 and -0.68 respectively, p<0.05) as were CIQ Home Integration (r=-0.36 and -0.38; p<0.05) and CIQ Social Integration (r=-0.46 and -0.38, p<0.05, TBI) (Kaplan, 2001).

Mail Administration: Using a mail questionnaire based on the modifications of Sander et al. (1999), ((Kuipers et al., 2004); ABI) reported an 80.2% completion rate for CIQ questionnaires by patients and 77.7% among proxy recipients. Home competency subscales had the highest completion rates in both groups, while social interaction had the lowest. Proxy scores on the home integration scale were significantly lower than patient scores (p=0.019). Item-to-total correlations ranged from 0.19 to 0.63 and subscale-to-total correlations were reported to be 0.73(home integration), 0.64 (social interaction) and 0.54 (productive activities). CIQ scores correlated with scores on the Sydney Psychosocial Re-integration Scale as follows (0.56 and 0.60 for patient and proxy scores, respectively): CIQ home competency correlated with Independent Living (0.42 and 0.57 for patient and proxy respectively), CIQ Social Interaction with Interpersonal Relationships (0.45 and 0.49 for patient and proxy), CIQ Productive Activity and Occupational Activity (0.42 and 0.41 for patient and proxy scores).

Agreement between scores derived from patient versus significant other telephone interviews was reported to be ICC=0.43 for home integration, 0.65 for social integration, and 0.81 for productivity subscales of the CIQ ((Tepper et al., 1996); TBI).

Agreement between patient and proxy scores obtained via telephone interview was reported to by 0.78 for CIQ total, 0.79 for home integration, 0.52 for social integration, and 0.84 for productivity subscales. Poorest agreements were noted for items that were most subjective and required opinion/judgement. In cognitive areas, proxies tended to score patients lower than the patients did themselves, while in activity areas, proxies tended to score patients higher than the patients themselves (Cusick et al., 2000).

Agreement between patient and proxy ranged from κ=0.43-0.70 on CIQ home integration subscale, 0.42-0.60 on the social integration subscale, and 0.69-0.94 on the productivity subscale. Significant differences were reported between patient and family member ratings on the home integration subscale (p<0.01) and total CIQ scores (p<0.05). In both cases, patient scores indicated higher levels of integration than scores derived from family member interviews. However, 80% of variance in total CIQ scores could be attributed to home integration sub-scores ((Sander et al., 1997); TBI).

When informants were interviewed, Willer et al. (1993) reported test-retest reliability of 0.97 for CIQ total scores, r=0.90 for social integration, 0.96 for home integration, and 0.97 for productivity subscales. Correlations between ratings provided by individuals with brain injury and family members were reported to be 0.81 for home integration, 0.74 for social integration and 0.96 for productivity, while total CIQ scores were also strongly correlated (r=0.89).

Family member and patient assessments were reported to be correlated, with r=0.81 for home integration, 0.74 for social integration, and 0.96 for productive activity (Willer et al., 1994).