How To (and How Not To) Assess the Integrity of Managers (19 pages)
Table 2 contains the means, standard deviations, and distributions for the five competency scales.
As expected, t tests indicated that the mean rating for Integrity was significantly higher than the
mean rating for every other dimension. The smallest difference occurred between Integrity
ϭ 2.68, SD ϭ .40) and Execution (M ϭ 2.31, SD ϭ .42) where the Integrity score was .90
standard deviations higher, paired t (671)
ϭ 22.88, p Ͻ .001.
Table 2 shows a second way to consider the elevation of the Integrity ratings. The percentages
in the right side of the table indicate the proportion of the total sample scoring between each value
on the five-point response scale. For instance, the proportion of the sample rated below the middle
ϭ effective,” was 21.1% for Vision, 19.3% for Execution, 25.3% for Managerial
Courage, 35.1% for Building Talent, but only 4.0% for Integrity. Not a single one of the 672
managers in the sample received an Integrity score below “1
ϭ adequate,” presumably the
minimally acceptable level. These results are consistent with our claim that competency ratings do
not identify managers with integrity issues.
Table 3 presents the correlations among the five competencies. As is typically the case with
competency ratings, and as was specified in the measurement model tested with CFA, the scales are
correlated with each other. However, Integrity was most highly correlated with Building Talent,
which reflects a manager’s concern for subordinates. This is consistent with the definitional position
that integrity involves sensitivity to the needs and rights of other people.
Finally, we examined the relationship between subordinates’ competency ratings and bosses’
overall performance rankings. First, we compared the average rating on each competency scale for
managers classified as “highest performing,” “solidly performing,” and “lowest performing.” The
results are presented in Table 4. Analysis of variance with follow-up Tukey tests for honestly
significant differences indicated that the means for Vision, Execution, and Managerial Courage were
significantly higher for the highest performing managers compared to the solidly- and lowest
performing managers ( p
Ͻ .001; partial eta
ϭ .04, .04, and .02, respectively). The means for
Building Talent and Integrity were not significantly different across the three levels of overall
performance. In other words, competency ratings of Integrity did not distinguish the high- from
low-performing managers. In fact, there was only a .11 scale-point difference between the average
Integrity rating for managers in the “highest performing” and “lowest performing” categories.
Further, the mean Integrity rating for “lowest performing” managers was 2.64, which is significantly
higher than the mean rating for each of the other four competencies for the “highest performing”
We also used the competency scores as a set of five covariates in a logit regression analysis to
predict overall performance. Logit regression is analogous to the familiar linear regression model,
except logit regression is more appropriate for criterion variables that are ordinal and categorical like
Descriptive Statistics and Distributions of Competency Scale Scores
Percentage of Sample Scoring between Scale Points
Managerial Courage 2.22 (.42)
2.68 (.40) 1.29 3.89
ϭ minimum; Max. ϭ maximum. Scale points: 0 ϭ Ineffective; 1 ϭ Adequate; 2 ϭ Effective; 3 ϭ
Very effective; 4
ϭ Extremely effective.
KAISER AND HOGAN