Scientific Validity Of the Leadership Matrix About the Measure of Practical Judgment
Two University of Pennsylvania psychologists
have completed the first of a series
of studies of the Leadership Matrix. The initial results indicate that its internal
consistency is excellent.
In an initial reliability test, Robert J. DeRubeis, Ph.D. and Michael Feeley, Ph.D. used
Cronbach's Coefficient Alpha and Spearman-Brown's split-hall method to determine the
instrument's internal consistency.
Reliability measures the consistency with which each factor is measuring the same thing,
content validity. Does the instrument measure what it purports to measure? (For example,
are the "S" adjectives actually measuring the personality construct Sociability as we define it?)
To measure content validity, a rating method is now used every time the Matrix is administered.
Those profiled were asked to assign a percentage of accuracy from 0 to 100% to the description
of their profile. The accuracy ratings collected to date average 92%.
The following are excerpts from DeRubeis' and Feeley's report:
- Measures the degree to which two separate variables are related
on a scale from -1.00 (perfect negative correlation) to +1.00 (perfect positive correlation.)
Between the two extremes is a zero (0) correlation, which is the equivalent of the flip of
a coin. For personality profiles like the MRA Leadership Matrix, positive correlations of +.80 or higher
is considered excellent.
1. Internal Consistency of the Entire Instrument
Coefficient Alpha for all items of the Matrix was found to be +.96. The split-half
reliability for all items also was +.96.
Internal consistency of the four factors (personality traits).
Ascendancy (A) - Coefficient Alpha = +.90
Split-half = +.80, rising to +.91 when corrected for the number of items of the scale.
Sociability (S) - Coefficient Alpha = +.91
Split-half = +.85, rising to +.91 when corrected for all items.
Emotional Adjustment (E) - Coefficient Alpha = +.85
Split-half = +.84 rising to +.91 when corrected for all items.
Readjustment (R) Coefficient Alpha = +.90
Split- half = +.84, rising to +.91 when corrected for all items.