Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
The assessment of work competencies and especially the development of valid and
reliable measures for their assessment have attracted limited attention among re-
searchers in the field of personnel psychology. On the contrary, human resources
practitioners have used the competency approach at length, especially during the
1980s and 1990s. Boyatzis (1982) first defined competencies as the certain charac-
teristics or abilities of the person that enable him or her to demonstrate the appro-
priate specific actions, thereby leading to effective work performance. Thus, on an
individual level, competencies are the capabilities the employee brings to the job.
Sparrow (1997), in a review of the use of organizational competencies in personnel
selection and assessment, defined competencies as those behavioral repertoires
(sets of behavioral patterns) that some people can carry out more effectively than
others, including all those behaviors that employees bring into the organization to
perform well.
Boam and Sparrow (1992), describing the major changes occurring in human
resources management, argued that organizations are looking for an approach
that will enable them to bring changes “by describing global issues in a way that
is sensitive to the local context … drawing upon the language of line managers
Requests for reprints should be sent to Ioannis Nikolaou, 39 Euripides St., Piraeus, Greece 185 32.
E-mail: inikol@aueb.gr
310 NIKOLAOU
rent study describes an attempt to develop and validate a measure of generic work
competencies.
METHOD
Study 1
Twenty-four managers participated providing useful data for 107 subordinates.
The majority of the managerial sample consisted of men (58%) between ages 35
GENERIC WORK COMPETENCIES 313
and 45 (67%) who were university graduates (46%). Table 1 shows the descriptive
results of the four scales of the competencies clusters and the job performance in-
dex along with their intercorrelations. For ease of comparison, the scale scores are
the average of the individual items of each scale.
Table 1 clearly shows that the competency scales are highly intercorrelated. The
first two clusters, the task and the action–leadership orientation, particularly have a
very high correlation coefficient of .92. Also, the job performance index has statis-
tically significant correlations with all four clusters. According to the managers of
the sample, a strong positive relation exists among all four types of work compe-
tencies and job performance, although the strength of this relation varies among
clusters of competencies.
The reliability of the competencies measure was very high. The Cronbach’s al-
pha was .98 and the Guttman’ s split-half was .97 (corrected for length). All the in-
dividual scales had also very satisfactory Cronbach’s alpha ranging from .92 up to
.96. Similar results came up for the job performance index (α = .95). The high in-
ternal consistency reliability is attributed to the high homogeneity and similarity of
the competency items.
To explore the adequacy of the sample for carrying out the factor analysis, the
Kaiser-Meyer-Olkin measure was estimated, reaching a value of .93, which is con-
sidered very acceptable (Norussis, 1994). Moreover, another index of the sample
suitability, the Bartlett’s test of sphericity, reached a statistically significant level.
Following Gorsuch’s (1983) recommendations, these are two basic requirements
to proceed with the exploratory factor analysis because they explore whether the
intercorrelation matrix is suitable for performing the factor analysis.
An initial exploratory factor analysis using principal components and oblimin
rotation was carried out. The extraction method of principal components was cho-
sen to extract linear combinations of the variables explaining the majority of the
variance (Kline, 1994), whereas oblimin rotation was applied because it was ex-
pected that the resulted factors would be highly intercorrelated. The latter was in-
deed the case as Table 1 shows. The results of the factor analysis yielded four fac-
tors explaining 76% of the variance; the people-orientation cluster and the
TABLE 1
Descriptives and Intercorrelations of the Scales (Study 1)
Scales N M SD α 1 2 3 4
communication skills orientation emerged very clearly but the remaining items
were distributed between the remaining two factors. The first point that should be
made here is that the participants could apparently distinguish the concepts of peo-
ple orientation and communication skills easier than the other two competencies
clusters (i.e., the task orientation and the action–leadership orientation). Subse-
quently, these two clear-cut factors (people and communication skills) were in-
cluded in the subsequent analyses and that the focus would turn to the interpreta-
tion of the remaining two factors.
Numerous exploratory factor analyses were conducted, repetitively excluding
items with low factor loadings, trying to identify a simple, coherent, and interpret-
able factor structure, but all the results were similar. Four factors were always com-
ing up; two of them were the people and communication factors, whereas the other
two were combinations of the task and action–leadership items. Subsequently, be-
cause the problem mainly concerned the task and action–leadership clusters, one
of the clusters was dropped because the participants could not distinguish between
these two groups of competencies. Thus, the action–leadership cluster was kept
and the task- orientation items were omitted. This was considered more appropri-
ate because it would be the only way of measuring leadership potential, which was
of more interest to the participant-managers, who also received feedback for their
responses, compared to the task-orientation competencies.
A hierarchical regression analysis (Cohen & Cohen, 1983) was also conducted to
examine whether the remaining items could predict the dropped items of the task-ori-
entation cluster. The three scales (people, communication, and action–leadership ori-
entation) were entered as independent variables in the equation with the task scale as
dependent. Table 2 summarizes the results of the regression analysis, which shows that
two of the remaining three scales (the action–leadership orientation and the communi-
cation skills) predict the task orientation very well with a multiple R = .932 (R2 = .869,
F(2,104) = 345.877, p < .000). Dropping the task-orientation items from the subse-
quent analyses should not be considered inappropriate because the remaining three
competencies clusters explained almost 87% of its variance.
The first exploratory factor analysis, using principal component analysis and
oblimin rotation, including all but the task-orientation items resulted again in a
4-factor solution, with the action–leadership items divided in two groups. But
when a 3-factor solution was requested, a very clear and easily identified solution
came out, explaining 72.237% of the total variance. Each competency scale
emerged very clearly corresponding to each of the three factors. The first factor,
where the action–leadership items were loaded, explained the majority of the vari-
ance (56.71%). To make the measure more efficient, a few more items, particularly
those with smaller loadings on the perspective factors, were omitted. Six items
were chosen for the action–leadership and the people-orientation clusters respec-
tively, and five for the communication skills cluster. Table 3 shows the resulting
factor pattern matrix, explaining a total variance of 79.5%.
TABLE 2
Hierarchical Regression Analysis of the Competencies Scales (Study 1)
1
Action–leadership .92 .84 .84 573.37 .00
2
Action–leadership .93 .87 .87 345.87 .00
communication
B SE of B β T Sig of T
1
Action–leadership .93 .04 .92 23.94 .00
Constant .71 .18 4.05 .00
2
Action–leadership .73 .06 .72 12.65 .00
Communication .29 .06 .25 4.37 .00
Constant .14 .20 0.70 .48
TABLE 3
Exploratory Factor Analysis of the Competencies Scale (Study 1)
Factor
Variables Items 1 2 3
Action–leadership Looks for new activities within his or her work .92 –.02 .00
environment
Takes the role of the leader in group activities .89 –.04 –.12
Looks for stimulation at work .84 .00 .12
Motivates his or her colleagues at work .81 –.05 .08
Is willing to commit himself or herself to new tasks .77 –.17 .02
Behaves dynamically at work .59 –.16 .21
People Displays kindness toward his or her colleagues .02 –.94 –.05
Shows consideration for his or her colleagues .08 –.92 –.05
Has good relationships with most of his or her colleagues .03 –.90 .03
Shows friendly behavior within the organization .04 –.86 .10
Respects his or her colleagues .16 –.85 –.07
Shows positive feelings toward his or her colleagues .04 –.77 .07
Communication Uses correct grammar in writing –.00 .02 .95
Uses correct spelling in writing .08 .02 .88
Uses an appropriate style in writing .30 .16 .78
Uses suitable language both in writing and speaking .41 .11 .67
Speaks clearly –.15 –.37 .67
315
316 NIKOLAOU
Following that, the relation of these three scales with job performance was ex-
plored. Thus, a hierarchical regression analysis was conducted to examine how
well the job performance index is predicted by the three competency clusters. Ta-
ble 4 summarizes the results of this regression analysis, which shows that the best
predictor of the job-performance index was the action–leadership competencies
cluster and not the other two clusters. The action–leadership was included in the
equation with a multiple R = .706 (R2 =. 499, F(1,105) = 104.629, p < .000). Subse-
quently, one can argue that the participants of the study do not consider good com-
munication skills or friendly behavior toward others within the organization as be-
ing associated with higher job proficiency, at least not as much as the competencies
included in the action–leadership orientation cluster.
Study 2
Following the development of the measure, we decided to conduct a second study
to examine the criterion-related validity of the new instrument. The sample con-
sisted of 57 managers, the majority of whom were men (72%) between the ages of
35 and 45 (53%) and were university graduates (49%). The participants were em-
ployed in 22 Greek organizations that completed the competencies and the overall
job performance measures for their subordinates (N = 218). The majority of the
firms operated in the services sector (72%) whereas 18% were industries and 10%
were commercial companies. Table 5 presents the descriptive characteristics and
the intercorrelations of the study’s variables, which all show satisfactory internal
consistency with action–leadership competencies showing the highest correlation
with job performance ratings.
An exploratory factor analysis of the competencies measure, using principal
component analysis and oblimin rotation, yielded results similar to the first study
establishing the construct validity of the measure (see Table 6). The three expected
factors emerged very clearly explaining a total variance of 84.9%. A hierarchical
regression analysis was also conducted with overall job performance as the de-
TABLE 4
Hierarchical Regression Analysis of the Three Competencies Scales
(Study 1)
B SE of B β T Sig of T
TABLE 5
Descriptives and Intercorrelations of the Variables (Study 2)
Scales N M SD α 1 2 3
TABLE 6
Exploratory Factor Analysis of the Competencies Scale (Study 2)
Factor
Variables Items 1 2 3
People Displays kindness toward his or her colleagues 1.02 –.01 –.06
Respects his or her colleagues .97 .08 –.09
Shows positive feelings toward his or her colleagues .93 .00 .02
Shows friendly behavior within the organization .91 .01 .03
Has god relationships with most of his or her .88 .01 .05
colleagues
Shows consideration for his or her colleagues .81 –.01 .16
Communication Uses correct spelling in writing .05 .96 –.09
Uses correct grammar in writing –.03 .95 .03
Uses suitable language both in writing and speaking .06 .92 –.02
Uses an appropriate style in writing –.04 .92 .08
Speaks clearly .10 .68 .14
Action–leadership Takes the role of the leader in group activities –.03 –.02 .93
Motivates his or her colleagues at work .05 –.08 .90
Behaves dynamically at work .02 .08 .81
Looks for new activities within his or her work .12 .03 .79
environment
Looks for stimulation at work –.04 .25 .74
Is willing to commit him or herself to new tasks .41 .03 .45
Note. Numbers in bold represent XXXXXXXXX. Extraction method: principal component anal-
ysis; Rotation method: oblimin with Kaiser normalization.
pendent variable and the three competency scales as the predictors, eliciting simi-
lar results with the results of the first study (see Table 7).
The results of these studies provide evidence for the construct and criterion-re-
lated validity of a new measure of generic work competencies. The measure cap-
tures three types of competencies: action–leadership, people orientation, and com-
munication skills. The existence of a valid and reliable measure of generic work
318 NIKOLAOU
TABLE 7
Hierarchical Regression Analysis of the Three Competencies Scales
(Study 2)
B SE of B β T Sig of T
competencies in Greece is a very important first step for the development of com-
petency management as part of Greek human resources management.
Our current methodology approached competencies in a “psychologi-
cal–psychometric” way in terms of conception and assessment to differentiate
from “broad” approaches in competencies in the field of human resources manage-
ment, which include in the definition of competencies concepts such as motives,
traits, attitudes or values, content knowledge, or cognitive or behavioral skills
(Hooghiemstra, 1992). Adapting this approach allowed us to build a simple-to-use
but simultaneously effective instrument both for academics researching compe-
tence at work as well as human resources professionals in performance and com-
petency management.
REFERENCES