among employees, determining training needs, setting goals, objectivity, and
overall ease of use. The trait scale was viewed by the Dutch managers as good or
as slightly better than the BARS.
Working in Israel, Tziner and Kopelman ( 1988 ) found that the use of BOS was
rated as signiWcantly higher by appraisers than were trait scales on three goal-
setting dimensions, namely, goal clarity, goal acceptance, and goal commitment.
These results were due to the behavioral speciWcity provided by the BOS. In
a subsequent study in Israel, appraisals based on BOS resulted in higher subse-
quent job satisfaction and organizational commitment than appraisals based on
trait scales (Tziner and Latham 1989 ). The speciWcity of the BOS strengthened
employees’ feelings of control over their work, and minimized feelings of ambi-
guity regarding expectations of them. Thus, both Tziner and Kopelman ( 2002 ), as
well as Bernardin ( 2005 ), concluded that BOS are preferable to BARS and
trait scales because they convey what the person ‘must do,’ feedback is perceived
as factual, objective, and unbiased; and the feedback is conducive to setting
speciWc challenging goals. SimilarWndings have been obtained with Canadian
managers regarding perceptions of fairness in the appraisal process (Latham and
Seijts 1997 ).
The superiority of the BOS to other appraisal scales likely reXects the fact that
they are based on Wherry and Bartlett’s ( 1982 ) theory of rating. Among the
theorems and corollaries are the following: ( 1 ) rating scales which assess behaviors
that are maximally controlled by the ratee lead to more accurate ratings than those
which refer to the tasks aVected primarily by the situation; ( 2 ) raters vary in the
accuracy of ratings given in direct proportion to the relevancy of their previous
contacts with the ratee; ( 3 ) rating items which refer to frequently performed acts
are rated more accurately than those which refer to acts performed rarely; ( 4 ) the
rater makes more accurate ratings when forewarned of the behaviors to be rated
because this focuses attention on the pertinent behaviors; ( 5 ) deliberate direction
of attention to the behaviors to be assessed reduces rating bias; ( 6 ) physical features
of a scale which facilitate recall of the actual perception of behavior increase rating
accuracy (e.g. larger descriptive behavioral statements are preferable to single value
words or simple phrases); ( 7 ) keeping a written record between rating periods of
speciWcally observed critical incidents improves the objectivity of recall; ( 8 ) stress-
ing the importance of appraisals to the organization or society as a whole decreases
bias and increases rater accuracy; ( 9 ) observation with intention to remember
facilitates recall; ( 10 ) behaviors which are readily classiWed by an observer within
a special category have smaller overall bias components; similarly, rating items that
are factorially unidimensional result in relatively less overall bias than items which
have a complex factor pattern.
In developing BOS, consistent with this theory, Latham and Wexley ( 1994 )
stressed that appraisers must be individuals who are aware of the aims and
objectives of an employee’s job, frequently observe the person’s behavior on the
performance management 369