How To (and How Not To) Assess the Integrity of Managers (19 pages)
The previous study indicated how ratings focused on positive behaviors may not identify
managers with integrity problems, but ratings focused on the undesirable portion of the domain may.
As Craig and Gustafson’s (1998) research led them to conclude, “global perceptions of supervisor
integrity . . . [are] . . . a function of discrete, and primarily destructive, supervisor behaviors” (p.
134, italics added).
Competency measures do not include items from the low end of the integrity continuum for at
least three reasons. First, leadership research traditionally emphasizes the positive qualities of
leaders (Kellerman, 2004; Padilla et al., 2007). Second, competencies are positive attributes that
characterize high performers (Boyatzis, 1982). This is consistent with the tendency of students of
leadership to seek “the right stuff” while ignoring “the wrong stuff.” A notable exception is research
studying managerial derailment (cf. McCall & Lombardo, 1983). A key lesson is that derailment is
less about lacking “the right stuff” and more about having “the wrong stuff” (e.g., arrogance and
abrasiveness; J. Hogan et al., 2010). Finally, it is more politically correct to ask if managers tell the
truth than to ask if they lie, cheat, and steal. But refusing to ask such pointed questions makes it hard
to identify managers who lack integrity.
It is difficult to rate low-integrity behaviors because they are hard to observe. For example,
employees are rarely caught stealing; however, the base rate of employee theft is about 50%
(Wimbush & Dalton, 1997). Although subordinates may not directly observe their manager’s
misconduct, they can estimate the likelihood that their manager would misbehave when opportu-
nities arise (Craig & Gustafson, 1998).
Moses, Hollenbeck, and Sorcher (1993) argued that data regarding a manager’s observed
behavior is less useful than data based on how others expect a manager to behave in the future. This
is because ratings reflect perceptions of behavior, not actual behavior. As Moses et al., (p. 286) note:
“In many cases, the behaviors behind others’ perceptions are never clearly delineated. The percep-
tions are formed as a result of cumulative observations and inferences about an individual.
Sometimes these inferences are based on concrete behavioral data. More often they are the collective
hypothesis which forms the aura behind a person’s reputation.” Moreover, a manager’s reputation—
that is, shared expectations about his or her future behavior—is the primary determinant of how
other people respond to that manager.
Asking subordinates to estimate the likelihood that a manager will behave unethically is
consistent with this view. People intuitively estimate the likelihood that others are cooperative or
selfish (Macy & Skvoretz, 1998). People search for signs of character in others— eye contact in
conversations, the company they keep—when deciding whether to trust them (Frank, 1988). Given
the vulnerability of their position, subordinates should be acutely alert to cues about a manager’s
trustworthiness (Van Vugt et al., 2008).
Craig and Gustafson (1998) developed and validated the Perceived Leader Integrity Scale
(PLIS), which is the epitome of the dubious reputation method. The scale contains 31 items focused
on unethical behaviors. The items reflect both character (e.g., “is vindictive”) and bad conduct (e.g.,
Summary of Logit Regression Analysis Using Competency Scales to Predict Overall
Nagelkerke Pseudo R
Wald parameters are analogous to beta-weights in linear regression.
KAISER AND HOGAN