Validation Procedure

To validate the individual scales in the CCI, we used two approaches.  First we used the traditional method of factor analysis to show uni-dimensionality combined with reliability analysis (Cronbach’s alpha) to show that each set of items was correlated with one latent construct. Next we conducted a study that involved correlating CCI scores with the evaluations of close friends and relatives of the respondents.  Both analyses are summarized below.

Factor and Reliability Analysis:

For the factor analysis, we examined the Scree tests, factor loadings and eigenvalues to determine whether each scale had only one dimension.  For the reliability analysis, we sought alphas at or near the 0.7 level, seeking higher alphas for scales that included more items.  The results of this analysis were as follows:

Kindness & Generosity (7 items)

Self-Other Evaluations:

For a preliminary test of construct validity, we distributed the CCI to a number of MBA students.   Simultaneously, we provided them with a supplemental "Friends and Family Form" which they were to give to "the person who knows them best."   The supplemental form is brief and simply asked the respondent "to what extent" the person they were rating was joyful, patient, faithful, etc.  We were careful to define the terms for the respondents to maximize consistent use of the form across raters.  The form is available for inspection by clicking here. Moreover, the entire process was anonymous.  We used code numbers to link CCI forms with Friend and Family Forms, and the latter forms were mailed to us confidentially with the postage paid by the researcher.

We received complete data for 61 individuals.  We then correlated the friend/family evaluations with the CCI scores.  Despite the relatively small sample size, the results indicated statistically significant correlations across almost all constructs, providing further evidence of the validity of the CCI.

The self-other correlations were as follows: