The Effects of Behaviorally Anchored Rating Scales as ...
Transcript of The Effects of Behaviorally Anchored Rating Scales as ...
Western Michigan University Western Michigan University
ScholarWorks at WMU ScholarWorks at WMU
Master's Theses Graduate College
12-1982
The Effects of Behaviorally Anchored Rating Scales as The Effects of Behaviorally Anchored Rating Scales as
Antecedent and Consequent Stimuli upon Instructional Behaviors Antecedent and Consequent Stimuli upon Instructional Behaviors
in a Special Education Setting in a Special Education Setting
Dennis L. Van Hartesvelt
Follow this and additional works at: https://scholarworks.wmich.edu/masters_theses
Part of the Industrial and Organizational Psychology Commons
Recommended Citation Recommended Citation Van Hartesvelt, Dennis L., "The Effects of Behaviorally Anchored Rating Scales as Antecedent and Consequent Stimuli upon Instructional Behaviors in a Special Education Setting" (1982). Master's Theses. 1748. https://scholarworks.wmich.edu/masters_theses/1748
This Masters Thesis-Open Access is brought to you for free and open access by the Graduate College at ScholarWorks at WMU. It has been accepted for inclusion in Master's Theses by an authorized administrator of ScholarWorks at WMU. For more information, please contact [email protected].
THE EFFECTS OF BEHAVIORALLY ANCHORED RATING SCALES AS ANTECEDENT AND CONSEQUENT STIMULI UPON INSTRUCTIONAL
BEHAVIORS IN A SPECIAL EDUCATION SETTING
by
Dennis L. Van H a rte sv e lt
A Thesis Submitted to the
F acu lty o f The Graduate C ollege in p a r t i a l fu lf i l lm e n t o f the
requirem ents fo r the Degree of M aster o f A rts Department o f Psychology
W estern Michigan U niversity Kalamazoo, Michigan
December 1982
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
THE EFFECTS OF BEHAVIORALLY ANCHORED RATING SCALES AS ANTECEDENT AND CONSEQUENT STIMULI UPON INSTRUCTIONAL
BEHAVIORS IN A SPECIAL EDUCATION SETTING
Dennis L. Van H a rte sv e lt, M. A.
Western Michigan U n iv ers ity , 1982
The p resen t study was an attem pt to determ ine the e f fe c ts of the
B ehaviorally Anchored R ating Scale (BARS) upon th ree d i s t in c t in s tru c
t io n a l behaviors when p resen ted befo re In s tru c tio n a l sessio n s and when
p resen ted a f te r In s tru c t io n a l se ss io n s . A m u ltip le b a se lin e design
across behaviors was used w ith fou r su b jec ts w ith a re v e rs a l element
used w ith one of those su b je c ts . A s l ig h t p o s it iv e e f f e c t of BARS
p re sen ta tio n upon in s tru c t io n a l behaviors occurred. No s ig n if ic a n t
d iffe ren ces were observed between p re in s tru e tio n a l and p o s tin s tru c tio n a l
BARS p re se n ta tio n . The re levance of th e p resen t study i s explored in
r e la t io n to previous BARS s tu d ie s , along w ith the im p lica tions of the
p re sen t study.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACKNOWLEDGEMENTS
There a re sev e ra l persons whose a id and cooperation were e s s e n tia l
to the completion of t h i s th e s is . I would f i r s t l ik e to thank the
ad m in is tra tio n a t the Croyden Avenue School o f the Kalamazoo V alley
In term ed ia te School D is t r ic t fo r providing th e research f a c i l i t i e s for
th i s p ro je c t. I would a lso l ik e to thank Joan Lundahl, Michele
Polver-Rueber, and Karin Uhlman fo r providing s ta f f and s tu den ts as
w e ll as Wayne Fuqua, Ph. D ., and Norman Pe te rson , Ph. D ., fo r providing
me w ith academic guidance which s e t the s tag e fo r th is th e s is . And
f in a l ly , my most s in ce re thanks a re o ffe red to Dale Brethower, Ph. D .,
who led me through the confusing and d i f f i c u l t ta sk of tu rn ing a few
id eas find academic preparedness in to a major l i t e r a r y work. W ithout
th i s h e lp , th i s ta sk would no t have been accomplished.
Dennis L. Van H arte sv e lt
i i
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
INFORMATION TO USERS
This reproduction was made from a copy of a document sent to us for microfilming. While the most advanced technology has been used to photograph and reproduce this document, the quality of the reproduction is heavily dependent upon the quality o f the material submitted.
The following explanation o f techniques is provided to help clarify markings or notations which may appear on this reproduction.
1. The sign or “ target” for pages apparently lacking from the document photographed is “Missing Page(s)” . If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure complete continuity.
2. When an image on the film is obliterated with a round black mark, it is an indication o f either blurred copy because of movement during exposure, duplicate copy, or copyrighted materials that should not have been filmed. For blurred pages, a good image o f the page can be found in the adjacent frame. If copyrighted materials were deleted, a target note will appear listing the pages in the adjacent frame.
3. When a map, drawing or chart, etc., is part of the material being photographed, a definite method o f “sectioning” the material has been followed. It is customary to begin filming at the upper left hand comer o f a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again—beginning below the first row and continuing on until complete.
4. For illustrations that cannot be satisfactorily reproduced by xerographic means, photographic prints can be purchased at additional cost and inserted into your xerographic copy. These prints are available upon request from the Dissertations Customer Services Department.
5. Some pages in any document may have indistinct print. In all cases the best available copy has been filmed.
UniversityMicrofilms
Internationa)300 N. Zeeb Road Ann Arbor. Ml 48106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1320047
VAN HARTESVELT, DENNIS LEE
THE EFFECTS OF BEHAVIORALLY ANCHORED RATING SCALES AS ANTECEDENT AND CONSEQUENT STIMULI UPON INSTRUCTIONAL BEHAVIORS IN A SPECIAL EDUCATION SETTING
WESTERN MICHIGAN UNIVERSITY M.A. 1982
University Microfilms
International 300 N. Zeeb Road. Ann Arbor. MI 48106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS . . . 11
LIST OF FIGURES . ................................. . iv
INTRODUCTION . . . . . . . . . .................... 1
METHOD . . . . . . . . . . 16
S u b j e c t s ........................ . . . . . • . . . . . . . 16
S e ttin g . . .................... * . . . . • • • • • • * 1 6
Scale D evelopm ent....................... . . . . . . . . . 18
Experim ental D e s ig n ......................... . . . • ......................... ' . .. 19
Data C o llec tio n . . ....................................... > . • • • ....................19
BARS P resen ta tio n . . . . . ......................... . . . . . . . . . 2 0
R e l ia b il i ty . . . . . . . . . . . . . . . • . . . . . . . • • 21
RESULTS . ..................... .... . . . . . . . . . . . . . . . 22
DISCUSSION . ......................... .... • .......................... • * * * 33
APPENDIX A. SD P resen ta tio n . . . . . . . . . . ........................ 43
APPENDIX B. Prompting . ....................... . . . . . . . 44
APPENDIX C. Consequation . .......................... . ............................ 45
APPENDIX D. Session M onitoring Form . . . . . . . . . . . . . . . 46
APPENDIX E. In te ra c tio n s between Observer and Subject . . . . . * 47
BIBLIOGRAPHY . . . . . . . . 49
i l l
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF FIGURES
1. In s tru c to r Consequation, Prompting, and S® P resen ta tio n and T heir E ffec ts upon S tudent Responses as a Function o f Antecedent BARS P re se n ta tio n fo r Subject A . . . . . . . . . . 23
2. In s tru c to r Consequation, Prompting, and S® P resen ta tio n and Their E ffec ts upon S tudent Responses as a Function o f Antecedent BARS P re se n ta tio n fo r Subject B . . . . . . . . . . 24
3. In s tru c to r Consequation, Prompting, and S& P resen ta tio n and Their E ffe c ts upon Student Responses as a Function of Antecedent BARS P re se n ta tio n fo r Subject C .. . . . . . . . . . 2 5
4 . In s tru c to r Consequation, Prompting, and S® P resen ta tio n and T heir E ffec ts upon Student Responses as a Function o f Antecedent BARS P re se n ta tio n fo r Subject D ............................. 26
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
INTRODUCTION
The purpose o f th is study was to determ ine on a b a s ic and em piri
c a l le v e l whether the B ehaviorally Anchored Rating Scale (BARS), which
i s norm ally used as a performance ev alu a tio n to o l, could a lso serve as
a t ra in in g to o l using p re sen ta tio n o f the sc a le s as feedback fo r appro
p r ia te in s tru c t io n a l behaviors.
Performance ev a lu a tio n o f employees w ith in o rgan iza tio n s in the
w estern world has become a prominent to p ic in contemporary management
and psychology jo u rn a ls (A llan & Rosenberg, 1978; Borman, 1978; Keeley,
1978) and in v i r tu a l ly every textbook dealing w ith in d u s tr ia l psychology
(Kane & Law ler, 1979; McCormick & T if f in , 1974; Wexley & Yuki, 1977).
An o u tlin e o f common su b jec tiv e performance ev a lu a tio n to o ls and tech
niques has been provided by Wexley and Yuki (1977).
An essay evalua tion i s the most open-ended o f the techniques. Here,
the su perv iso r simply w r ite s , in essay form, a summary o f the perform
ance o f each subord inate. This form may be more personal than o th e rs ,
b u t i t i s time consuming and lack s q u a n tif ic a tio n fo r employee com
parison . In a group a p p ra is a l , the su perv iso r o f the employees to
be evaluated as w ell as th re e o r four o th e r su p erv iso rs fa m ilia r w ith
them meet to d iscu ss th e i r performance. This method tends to elim
in a te much e r ro r in the essay ev a lua tion ; but again , employees can
not be compared q u a n tita tiv e ly ; and i t i s very , very c o s tly to the
o rg an iza tio n . In a f i e ld review , a tra in e d personnel manager goes
in to various sec tio n s o r d iv is io n s o f the o rg an iza tio n to d iscu ss the
performance o f in d iv id u a l employees w ith su p erv iso rs and sometimes
1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
w ith the employees them selves. With the inform ation gained, the per
sonnel s p e c ia l i s t then w rite s an ev a lu a tio n o f the in d iv id u a ls ' per
formance which i s amended by th e i r s u p e rv iso r (s ) . This method cu ts
down on paperwork by sup erv iso rs and i s probably more c o n s is te n t, i f
n o t more accurate , than essay evalu a tio n s and group a p p ra is a ls . But
once again , employees cannot be q u a n tita tiv e ly compared; and i t i s pro
bably no t as accu ra te as observation-based ev a lu a tio n procedures.
Employee comparisons involve weighing up one employee a g a in s t
another to determ ine performance rank ings. A lte rn a tio n ranking i s the
sim plest o f the th ree comparison techniques, bu t i t may lack accuracy.
Here, the su p erv iso r determ ines the h ig h est and low est perform ing
employees in a p a r t ic u la r group and p laces th e i r names a t the h ig h est
and low est p o in ts on the sca le re sp e c tiv e ly . Then, he takes the next
h ig h est and low est perform ers and does the same. This i s repeated
u n t i l the l i s t is complete. In the p a ired comparison technique, each
employee is compared to every o th e r . This can be done on a m atrix
where the name of the b e t te r perform er i s in se r te d in the in te rs e c t io n
fo r the names o f each p a ir of employees. Rank can then be determined
by simply counting the number of times each employee was p re fe rre d .
While n o t performance s p e c i f ic , th is i s a f a i r ly accu ra te means of
ranking employees and can be used i f only rankings a re necessary . A
forced d is t r ib u t io n can be used fo r la rg e groups o f employees where
rough d isc rim in a tio n s between them w il l s u f f ic e and when i t can be
assumed th a t job performance conforms to a normal curve. Five spaces
a re provided in which the su p erv iso r p laces the names o f employees who
perform: (a) in the low est 10%; (b) in the next 20%; (c) in the middle
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40%; (d) in the nex t h ig h es t 20%; and (e) in the h ig h es t 10%. The prob
lem here is th a t job performance o ften does no t conform to norm ality ,
e sp e c ia lly where th e re a re e f fe c t iv e s e le c tio n and/or tra in in g techn i
ques ( th is would skew th e d is t r ib u t io n to the l e f t ) . And w hile employees
f i t in to n ea t l i t t l e c a te g o rie s , no inform ation i s gained about perform
ance le v e ls so sa la ry and promotion dec is io n s a re d i f f i c u l t to make
w ith th is comparative procedure.
The c r i t i c a l in c id e n t technique f i r s t described by Flanagan (1949)
i s an observation-based technique. The su perv iso r keeps a running log
fo r .each employee in which he records s p e c if ic in s tan ces o f c r i t i c a l
co n trib u to ry and o b s tru c tiv e work in c id e n ts . This log then i s re fe rre d
to d ire c t ly during the performance ev a lu a tio n . While th is i s an
observation-based technique, i t i s not q u a n tif ia b le and does no t take
in to account the normal ongoing behaviors by employees v i t a l l y impor
ta n t in a tta in in g e f fe c t iv e p roduction le v e ls .
The weighted c h e c k lis t i s one where s tatem ents about poor to
e x ce lle n t performance a re l i s t e d in random o rd er. The score values
a re undisclosed to the r a t e r who checks every statem ent on each l i s t
th a t rep re se n ts the em ployee's performance (various dimensions o f per
formance a reas and a t t i tu d e s a re o fte n in c lu d ed ), and the l i s t s a re
scored by a judge.. The s ta tem en ts t hemselves, however, o f te n give
away th e i r v a lues s in ce the r a te r s recognize various performance le v e ls
and those r a te r s may avoid checking negative s ta tem en ts . In a forced
choice c h e c k lis t , blocks o f four s tatem ents about poor to ex ce lle n t
performance a re l i s t e d where the r a te r checks one statem ent which i s
most l ik e the employee and one statem ent which i s l e a s t l ik e the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
employee. The statem ents a re then scored and a performance ra t in g can
be made. While p o ssib ly reducing r a t e r e r ro r , l i t t l e in form ation i s
gained about sp e c if ic performance to apply to personnel development.
Three ra t in g sca le s were o u tlin e d , two o f which were very s im ila r .
In the graphic ra tin g s c a le , the subheadings "poor," "average ," and
"ex c e lle n t" a re placed a t th e i r re sp ec tiv e p laces from l e f t to r ig h t
on the s c a le s , each of which has some performance o r a t t i tu d e heading.
The r a te r can then mark any p o in t on the sca le ( i . e . , between average
and e x ce llen t) to re p re sen t the em ployee's perform ance. The m ultis tep
sca le i s very s im ila r , except th a t the r a te r can mark only in the
th ree to f iv e evenly spaced boxes provided to re p re sen t from poor to
ex ce lle n t performance o r a t t i tu d e . The em ployee's dimension ra tin g s
can be compared w ith o th e r employees' ra tin g s o r can be averaged and
compared. Wexley and Yuki contend th a t both of these sc a le s a re q u ite
su scep tib le to r a te r e r ro r s .
The B ehaviorally Anchored S a ting Scale has probably been the most
widely d iscussed performance ev a lu a tio n technique in the l a s t two
decades. In th is Scale, a dimension o f performance i s ra te d on a
sca le which rep resen ts poor to e x c e lle n t performance by s p e c if ic
examples o f behavior. This i s sometimes c a lle d a Behavior Expectation
Scale because the anchors (behav ioral examples) a re o ften preceded by
"The employee can be expected to . . . . " The dimensions a re u su a lly
rep resen ted by th ree to ten anchors spread o u t over the continuum of
the s c a le . Each Anchor i s assigned a p a r t ic u la r num erical value th a t
i s meaningful only in re fe ren ce to the e n t i r e sca le which may be sca led
by ten th s o r by wholes and whose upper l im it may be from 2.5 to 10.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
At the time o f the performance ev a lu a tio n , the r a t e r checks the anchor
in each o f about f iv e to ten dimensions which most rep resen ts the
employee's ty p ic a l work perform ance. The performance ev alu a tio n can
be made num erically so th a t comparisons between employees can be done
along dimensions and along o v e ra ll averages. The corresponding
anchors can a lso add valuab le performance inform ation fo r personnel
development. As a r e s u l t o f th is f l e x i b i l i t y , the B ehaviorally
Anchored Rating Scale may be a b e t te r ev a lu a tio n to o l than the o th e rs .
However, w ith a l l these ev a lu a tio n to o ls a t hand, W illiams (1972)
claim s th a t th e re i s su b s ta n tia l evidence around the world to suggest
th a t many performance a p p ra isa ls a re a w aste of time. He provides
evidence th a t many a p p ra isa ls a re in accu ra te ; and in many cases,
whether the ap p ra isa ls a re accu ra te o r n o t, the inform ation i s no t
used as a b a s is fo r s a la ry r a is e s , promotion, o r personnel development.
W illiam s, however, has n o t given up on personnel ev a lu a tio n ; and he
promotes fu r th e r development o f e f fe c t iv e performance ev a lu a tio n .
The B ehaviorally Anchored Rating S ca le , which i s the focus of
th is study , has been tou ted as a prom ising new to o l in s t a f f and per
sonnel ev a lu a tio n , a t l e a s t on a s t a t i s t i c a l le v e l (Kearney, 1976).
Bernardin and Smith (1981) m ain tain th a t the B ehaviorally Anchored
Rating Scale (BARS) i s d i f f e r e n t from o th e r methods o f ev alu a tio n
q u a li ta t iv e ly in th a t i t i s designed to s tan d a rd ize the ra tin g and
the observation o f employees. They p o in t o u t th a t w ith standard ized
o b serv atio n , th e re i s a common ground fo r r e l ia b le ra tin g o f perform
ance.
The s p e c if ic methods fo r BARS development vary somewhat, b u t the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
author has encountered from f iv e to e ig h t s p e c if ic s tep s o f BARS
development in a nearly exhaustive skimming o f the BARS l i t e r a tu r e .
Blood (1974) has o u tlin e d th ree general s tep s which cover the ty p ic a l
BARS development. F i r s t , behav ioral in c id e n ts o f various degrees o f
job performance dimensions a re generated by a pool of people knowledgea
b le in the job fo r which the BARS i s being developed. Secondly, the
in c id en ts a re re tra n s la te d . This involves re a s so r tin g the in c id en ts
from a randomized pool in to th e i r proper dimensions. In o rder fo r
in c id en ts o r anchors to be re ta in e d , they must be r e l ia b ly p laced in
th e ir dimensions by a sound m ajo rity of the judges. F in a lly , the
anchors a re assigned s c a le values according to the le v e l o f perform
ance they rep re se n t. The mean of a l l values assigned i s the f in a l
sca le value o f the anchor, and in c id en ts w ith high variances a re d is
carded. O ccasionally , th e job dimensions themselves a re developed
through a generation and re tr a n s la t io n procedure where the dimensions
are reassigned from a random pool to groups o f anchors. There a re
numerous cookbook approaches in the l i t e r a t u r e which deal w ith BARS
development and which may be consu lted fo r fu r th e r d e ta i l s (B ernardin,
1977; B ernardin , A lvarez, & Cranny, 1976; B ernardin & Smith, 1981;
Borman & Dunnette, 1975; Borman & V allon, 1974; Campbell, Dunnette,
Arvey, & H e lle rv ik , 1973; Das, F ro s t, & Barnowe, 1979; D eC otiis, 1977;
D eCotiis, 1978; Dickinson & Z e llin g e r , 1980; Ivancevich , 1980b; Landy
& Guion, 1970; Kearney, 1976; Keavney & McGann, 1975; M illa rd , Luthans,
& Ottemann, 1976; Shapira & Shirom, 1980; Smith & K endall, 1963;
S tap les & Locander, 1972; Zedick & Baker, 1972; Zedeck, K afry, &
Jacobs, 1978).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The ty p ic a l performance ev alu a tio n w ithout a BARS u su a lly involves
a few sim ple s te p s . For hourly employees, th e i r sup erv iso rs decide how
w ell they perform th e i r s p e c if ic tasks and what th e i r work a t t i tu d e is
l ik e , and check th e ir a ttendance and p u n c tu a lity . For s a la r ie d
employees, a few o b je c tiv e s a re s e t a t the beginning o f the year such
as in an MBO program; and a t the annual ev a lu a tio n , some evalua tion
to o l may be used to determ ine o v e ra ll job performance, and o b jec tiv es
a re examined to see i f they have been met. The BARS appears to be
a su p erio r procedure to these on four p o in ts . F i r s t , th e superv iso rs
who use i t a re , in most c ase s , deeply involved in i t s development
which may provide some face o r con ten t v a l id i ty as w e ll as concurrent
v a l id i ty . Secondly, those sup erv iso rs a re thoroughly fa m ilia r w ith
the BARS' conceptions and param eters so they a re ex p erts in i t s use
which may provide some r e l i a b i l i t y in observation and ev a lu a tio n .
T hird , the sc a le dimensions a re broken down in to behav io ra l d esc rip
tio n s o f most o r a l l performance le v e ls (on a continuum o f poor to
ex ce llen t) as opposed to a one to f iv e r a t in g o f how the r a t e r be liev es
the ra te e i s doing in a g re a t number of a t t i tu d e s and performance
requirem ents. This provides o b je c tiv e s fo r the employee and d e f in i
tio n s of performance which should help them to avoid undesirab le
behaviors, th e re fo re ,le av in g room fo r d e s ired performance. F in a lly ,
the s p e c i f ic i ty of the to o l may make ev a lu a tio n e a s ie r thus promoting
more frequen t evalu a tio n . Wexley and Yuki (1977) p o in t out th a t f r e
quent evalua tions o f performance may help o rg an iza tio n s meet th e ir
s h o r t- and long-term o b je c tiv e s .
There i s evidence in psychometric terms to support the su p e r io r ity
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
of BARS over ocher ra t in g sc a le s , such as the Graphic Rating Scale,
the Mixed Standard Scale, and the M ultistep S cale . The BARS has been
shown to in c rease in te r r a te r r e l i a b i l i t y and to reduce r a te r b ia s
(Landy &Guion, 1970; M illard e t a l . , 1976; Zedeck & Baker, 1972) and
has decreased len iency and ra t in g v a r ia b i l i ty (Das e t a l . , 1979;
Shapira & Shirom, 1980; Vance, Kuhnert, & F a rr , 1978). Ivancevich
(1980b) re p o rts th a t w ith the im plem entation o f BARS fo r p ro fess io n a l
employees, th e re was an improvement in a t t i tu d e about equ ity in job
requirem ents and c la r i ty in feedback.
BARS has been used in a v a r ie ty o f jobs inc lud ing nurses (Smith
& K endall, 1963; Zedeck & Baker, 1972), engineers (Landy & Guion,
1970), r e t a i l personnel (Campbell e t a l . , 1973; S tap les & Locander,
1972), tank crews in the I s r a e l i army (Shapira & Shirom, 1980), co lleg e
p ro fesso rs (Das e t a l . , 1979; Dickinson & Z e llin g e r , 1980; Vance e t
a l . , 1978), bank managers (Schwind, 1978), and s a le s personnel
(Ivancevich, 1980a).
There a re , however, In c reasin g numbers o f s tu d ie s which s t a t e
th a t the BARS i s no b e t te r b u t no worse than o th e r form ats, such as
graphic ra tin g sca le s (Keaveny & McGann, 1975), behavior observation
Scales (Latham, Fay, & S a a ri, 1979) and mixed standard sca le s (Dickinson &
Z e llin g e r, 1980). Keeley (1978) claim s th a t the BARS i s lim ited in
a p p lic a tio n to ro u tin e ta sk s s in ce i t i s anchored to s p e c if ic in s tan ces
o f behavior and i t may no t be as f le x ib le as a g raphic ra tin g s c a le .
A tkin and Conlon (1978) o u tlin e o th e r p o ss ib le problems w ith the
BARS. F i r s t , th e re may be overlap o r nonexclusiveness o f item s w ith in
each dimension. Second, c r i t i c a l in c id en ts o f behavior (extrem ely
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
good o r poor performance) may stand o u t and be re f le c te d In the ra tin g
process w hile ro u tin e performance I s overlooked. And f in a l ly , th e pre
d isp o s it io n o f the r a te r toward ra te e s may a f f e c t h is ra tin g s o f those
people. But these problems would be common to a l l r a t in g procedures
and n e ith e r unique to the BARS, nor solved by the BARS.
Kingstrom and Bass (1981) c ite d sev e ra l s tu d ie s which c o n tra d ic t
the conclusions o f h e re to fo re popular pro-BARS s tu d ie s . In th e i r
review, Kingstrom and Bass show approxim ately equal numbers o f s tu d ie s
find ing su p e r io r ity w ith the BARS and w ith the a l te rn a te form at sca le s
(g raph ic , stimulated, mixed standard) in len iency , in te r r a t e r r e l i a b i l i t y ,
ra te e d is c r im in a b ili ty , dimension independence (h a lo ), and v a l id i ty .
A few s tu d ie s showed no d iffe ren c es in th e se . The a u th o rs ' conclusions
were th a t th e re were n e g lig ib le d iffe ren c es between BARS and a l te rn a te
form at s c a le s . This was q u a lif ie d , however, w ith th e f a c t th a t most
o f the a l te rn a te form ats used fo r comparison w ith BARS in th ese s tu d ie s
were developed from inform ation derived from BARS developmental proce
dures. This may in d ic a te th a t the s c a le development procedure and not
the s c a le i t s e l f i s the c r i t i c a l v a r ia b le .
Jacobs, K afry, and Zedeck (1980) concur w ith the find ings of
Kingstrom and Bass (1981). In a review o f a t l e a s t 39 s tu d ie s dealing
d ire c t ly w ith BARS comparisons, i t was th e i r conclusion th a t i t was
n e ith e r su p erio r nor in f e r io r to any o th e r form at i t has been compared
to . I t seems th a t the problems w ith BARS as a ra t in g to o l a re the
same as w ith any o th e r r a t in g to o l and th a t each may be j u s t as good
o r bad as the o th e r.
In shedding new l i g h t on the advantages o f and problems w ith BARS,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
the author p o in ts out th a t a l l these s tu d ie s lacked any raw an a ly s is
of the e f fe c ts o f BARS a p p lic a tio n upon employee performance. Wexley
and Yuki (1977) contend th a t " c u rre n t resea rch suggests th a t the
ex tra time and e f f o r t req u ired to b u ild behavior-based ra tin g sc a le s
may n o t be w arranted i f used so le ly fo r performance ra tin g s" (p . 207).
Blood (1974) has o u tlin ed fou r sp in o ffs fo r BARS development. F i r s t ,
as a lluded to e a r l i e r , the o rg an iza tio n can b e n e f it from a s c a le which
can be developed by re p re se n ta tiv e s o f va rio u s o rg an iza tio n a l le v e ls ,
th e re fo re , ga in ing im portant p e rsp ec tiv es o f the job fo r which the
BARS i s being developed. Secondly, the inform ation about the job
obtained in BARS development can be used to c re a te a v a lid tra in in g pro
gram. The tra in e e s would le a rn what i s expected o f them and how they
would be evaluated . T h ird ly , when th e re i s disagreem ent in the BARS
development over the v a lue o f anchors o r in what dimension they belong,
th e re i s a s ig n a l th a t o rg a n iz a tio n a l p o licy does not e x is t fo r these
performance requirem ents o r i t i s n o t being conveyed. I f th e re a re
problems w ith p o licy , they can then be ironed ou t according to the
inform ation gained in th e BARS development p rocess . Where p o licy
e x is ts but disagreem ents occur, the fo u rth sp in o ff comes in to p lay .
The o rg an iza tio n can determ ine where the breakdown i s in communication
and deal w ith i t .
In an endeavor to t e s t the param eters o f BARS a p p lic a tio n , the
p re sen t study was an a ttem pt on a sim ple, b a s ic , and em pirica l le v e l
to determ ine whether the BARS method, used in a very lim ited co n tex t,
would be fu n c tio n a lly r e la te d to performance improvements in terms of
a c tu a l behavior change as a r e s u l t o f BARS p re sen ta tio n s a lone. The
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
BARS inform ation provided by ra tin g s accompanied by sp e c if ic examples
o f performance can be likened to feedback. Feedback i s defined as
inform ation about performance, compared to a s tandard , and used to
guide performance. The inform ation about performance here would be
the performance ra t in g . The standard would be the h ig h e s t anchor, and
an a tta inm ent o f th is would dem onstrate a guidance o f performance.
Prue and Fairbank (1981) o u tl in e param eters of feedback im portant
in a t ta in in g optimum e ffe c tiv en e ss in the p a r t ic u la r s i tu a t io n where
i t i s to be used. They p o in t ou t th a t performance feedback i s popular
in attem pting to c re a te performance improvements and th a t an ex tensive
da ta base has proven th a t provid ing inform ation to groups and to
in d iv id u a ls about the q u an tity and q u a lity of th e i r performance does
improve i t . Feedback has c e r ta in advantages which enhance i t s popu
l a r i t y . I t i s inexpensive to d e liv e r ; i t i s easy to implement s in ce
e lab o ra te tra in in g i s n o t a p re re q u is ite to accu ra te ad m in is tra tio n
( in the case o f BARS, lim ited a p p lic a tio n may make i t s ex tensive
development p ro h ib i t iv e ) ; i t cu ts back the use o f nonpolicy av ersiv e
co n tro l procedures; and i t i s u se fu l where monetary rewards cannot be
delivered , as in a union environment.
Four feedback mechanisms have been o u tlin e d : m echanical,
se lf-re co rd ed , v e rb a l, and w r it te n . Mechanical feedback i s th a t pro
vided by an apparatus, such as a cummulative ta l le y odometer o r a
v ideotape o f w idget assembly. This i s an inexpensive form of feedback
sin ce i t has moderate i n i t i a l and l i t t l e ongoing c o s t and i t w i l l never
adm in ister an em otional reprimand. S elf-reco rded feedback i s achieved
by the employee recording h is own performance on an odometer o r by
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
w ritte n t a l ly . This i s advantageous when the employee has l i t t l e
superv ision o r when th e re i s no permanent p roduct. Verbal feedback
is th a t provided by a superv iso r to an employee on a d i r e c t in te r
personal le v e l . This i s an easy method o f feedback d e liv e ry ; b u t i t
may lack consistency , s in c e r i ty may no t be ev id en t, and i f an adverse
re la tio n sh ip e x is ts between a su p erv iso r and employee, i t may no t be
a t a l l e f fe c t iv e . W ritten feedback can e a s ily be made c o n s is te n t.
I t i s a permanent p roduct, and i t can e a s ily f i t in w ith any informa
tio n p rocessing system the o rg an iza tio n p re sen tly u t i l i z e s . The
inform ation gained from BARS as a feedback to o l a lso has these q u a li
t ie s .
The con ten t o f the feedback can inc lude in d iv id u a l performance
as a percentage o f the g roup 's performance, comparison of a g roup 's
performance w ith a s tandard , comparison o f a g roup 's performance w ith
previous performance, comparison o f a in d iv id u a l 's performance w ith
a standard , and comparison o f an in d iv id u a l 's performance w ith previous
performance. The inform ation contained in an in d iv id u a l BARS p resen ta
tio n can be compared w ith a s tandard (th e h ig h es t anchor) and w ith
previous performance (previous ran g e ).
Prue and Fairbank a lso d iscussed the e f fe c ts o f feedback upon
performance. They p o in t o u t th a t feedback e f fe c ts have in the p a s t
alm ost b lin d ly been ch a rac te rized as re in fo rc in g . S ince response
ra te s were increased , th is could be tru e ; b u t they a lso p o in t o u t th a t
o th e r co n sid era tio n s must be made. One of these would be the stim ulus
co n tro l which feedback ex e rts over behavior, such as when an organiza
tio n s e ts new requirem ents fo r re inforcem ent. Performance could be
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13ru le governed where feedback i s used fo r the f i r s t time in the o rgan i
za tio n so th e re would be no h is to ry o f con tingen t reinforcem ent.
Feedback could a lso a l t e r the o rg an iza tio n a l environment such th a t
inform al d iscu ssio n s would cen te r around performance improvement and
nonprogrammed so c ia l re in fo rc e rs would be a v a ila b le . I t could a lso
s e t up an atmosphere of com petition . F in a lly , ' they p o in t out th a t
w ith consequation based on o b jec tiv e d a ta , employees are more s a t i s f ie d
w ith th e i r work and surroundings which could enhance the feedback
system.
The e ffe c tiv en e ss o f BABS in improving performance was evaluated
in the p re sen t study by examining i t s e f fe c ts upon s t a f f in s tru c t io n a l
behaviors in a behavior m odifica tion o rien ted s p e c ia l education f a c i l i t y .
In the same f a c i l i t y , Shook, Johnson, and Uhlman (1978), in two sep ara te
experim ents, s tud ied graphing by s t a f f using (a) response e f f o r t
red u c tio n , in s tru c t io n s , and group feedback (a n o tic e board w ith a p er
centage of u p -to -d a te graphs l i s t e d ) , followed by (b) in s tru c tio n s ,
group feedback, in d iv id u a l feedback (n o tice o f percen t of d a ily graphs
which were c u r re n t) , and reinforcem ent (s o c ia l p r a is e ) . The l a t t e r
procedure appeared to be most e f fe c t iv e and most c o s tly , w hile
in d iv id u a l feedback alone appeared to be most co st e f fe c t iv e in terms
of response e f f o r t by experim enters to d e liv e r d e sc r ip tiv e p ra ise to
each su b jec t compared w ith one e a s i ly assembled c h a rt.
In an attem pt to analyze the feedback e f fe c ts o f the BARS, the
sca le in the p resen t study was presented as both an an tecedent and as
a consequence. Krumhus and M alott (1980) s tud ied th e e f fe c ts o f model
l in g , feedback before a se ss io n , and feedback a f t e r a session upon
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
d e sc rip tiv e so c ia l reinforcem ent of s tu den ts by tu to r s . No d iffe ren c e
was observed between immediate ( ju s t a f te r session ) and delayed (p r io r
to next session) feedback, but th e re were marked in c rease s from base
l in e to in te rv e n tio n in both p re - and p o s tse ss io n feedback and some
in c rease from p o s tse ss io n m odelling to p re - and p o s tse ss io n feedback
in tu to r responses. Feedback, both befo re and a f t e r se ss io n s , con
s is te d of a review o f a v ideotape o f the s u b je c t 's most re cen t sessio n
by the observer and the su b je c t, w ith the su b jec t c i t in g th re e in s tan ces
and noninstances of ap p ro p ria te d e sc r ip tiv e so c ia l reinforcem ent.
In a study by Sepler and Meyers (1978), " . . . verba l in s tru c
t io n alone was shown to be in s u f f ic ie n t fo r teach ing the f u l l range
o f behavior m od ifica tion s k i l l s requ ired to fu n c tio n e f fe c t iv e ly in
app lied s e t t in g s " (p. 198). In th is s tudy , p a rap ro fess io n a ls were
given a w ritte n p r e te s t , in s tru c te d in behavior m odifica tion s k i l l s ,
and then given a w ritte n p o s t te s t . They were a lso observed fo r appro
p r ia te responses throughout the study.
T rain ing was found to be e s s e n t ia l in in c reas in g the p roportion
o f s p e c if ic tra in in g behaviors by p a rap ro fe ss io n a ls during therapy
periods (Fabry & Reid, 1978) and in in c reas in g ap p ro p ria te S ^ 's
w hile reducing the elapsed time of each ta sk (Carnine & Fink, 1978).
As suggested by Blood (1974), the BARS could serve f i r s t as an
ev alua tion to o l to determ ine what s k i l l s a th e ra p is t dem onstrates and
lacks and secondly as a t ra in in g to o l by showing the th e ra p is t in
what way he o r she i s in e f fe c t iv e and what must be done to a t ta in
e ffe c t iv e n e s s . A t r a in e r could a lso use th is inform ation to d if fe re n
t i a l l y re in fo rce behaviors o r ap p ro p ria te approxim ations th e reo f
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
' 15
which c o n trib u te to e ffe c tiv e n e s s . As poin ted out above, such t r a in
ing was found to be im portant by Fabry and Reid (1978) and Carnine and
Fink (1978) in developing behavior m odifica tion s k i l l s . By surpassing
in s tru c t io n a lone , which Sepler and Myers (1978) found in s u f f ic ie n t
in tra in in g behavior m od ifica tion s k i l l s , the BARS could w ell be a
very e f f ic ie n t to o l fo r personnel development.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
METHOD
Subjects
The su b jec ts were a l l employees o f a sp e c ia l education f a c i l i t y
and were a l l fem ale. Subject A was a 24-year-o ld c e r t i f ie d teacher
who, in th is study, worked w ith th ree s tu d en ts from her classroom on
D istar Math. She had previous experience in D is ta r Math. Subject B
was a 21-year-o ld in s tru c t io n a l a id e working w ith the same th ree s tu
dents as Subject A in D is ta r Math. Subject B had no previous experience
in D ls ta r Math. Subjects A and B a lte rn a te d p ro g ressive le sso n s so
th a t n e ith e r ran the same le sso n , although le sso n s o ften repeated
d r i l l s from previous le sso n s . Subject C was a 28-year-old c e r t i f ie d
teacher who was working w ith the same th ree s tu d en ts as above whom
she did no t o therw ise normally work w ith . In a d d itio n , she worked
w ith one o f h e r own studen ts who jo in ed the group o f the o th e r th ree
in D is ta r Language. Subject C had minimal prev ious experience in
D is tar Language. Subject D was a 21-year-o ld p a rt- tim e parap ro fes-
s io n a l working w ith a d if f e r e n t s tu d en t than those above on D ista r
Math. She had no previous experience in D is ta r Math. A ll o f the
math lessons fo r Subjects A, B, and D were conducted from D is ta r Math
Book I . A. A ll o f the language le sso n s fo r S ub jec t C were conducted
from D is ta r Language Book I . A.
S e ttin g
Subjects A, B, and C a l l worked in the same a rea a t d if f e r e n t
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
tidies. The area included a long ta b le , on e i th e r s id e of which s a t
the s tu d en ts and the su b jec ts ( in s tru c to r s ) ; a blackboard behind the
in s tru c to r ; and D is ta r Language o r Math m a te ria ls on the ta b le .
S ubject D worked in an in d iv id u a l sess io n bo th , which included a
s tu d e n t 's desk, a c h a ir fac ing the s tu d en t in which the su b jec t s a t
w ith D is ta r Math m a te ria ls on the desk. The observer always s a t in
a p o s itio n somewhere behind the s tu d en t o r s tu d en ts so as not to
in te r f e re v is u a lly w ith the su b je c ts .
The s tu d en ts a l l a ttended classrooms where behavior m odifica tion
was the main mode of s tu d en t m otivation . This behavior o r ie n ta t io n
requ ired c o n s is te n t behaviors by the in s tru c to rs in p re ta sk o r an te
cedent stim ulus p re sen ta tio n , c o rre c tio n , o r prompting and p o s tta sk
o r consequent stim ulus p re se n ta tio n . P a rap ro fessio n a ls and in s tru c
tio n a l a ides such as Subjects B and D norm ally received in -s e rv ic e
tra in in g in classroom behavior management inc lud ing behavior m odifica
tio n in in s tru c tio n a l se ss io n s . They were u su a lly monitored bi-m onthly
in in s tru c t io n a l behavior m od ifica tion . C e r tif ie d teachers such as
Subjects A and C re g u la r ly p a r tic ip a te d in classroom behavior manage
ment conferences; bu t in the f a c i l i t y where they were employed during
th is study , Subjects A and C received no formal tra in in g nor were
they monitored in in s tru c t io n a l behavior m o d ifica tio n . In the p a s t,
however, they had both received in te n s iv e tra in in g in behavior modifi
ca tio n in in s tru c t io n a l sessio n s concurren t w ith previous p o s itio n s
which requ ired th is t ra in in g .
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
S cale Development
The th ree BABS sc a le s were developed by the au thor (see Appendices A-C)
from job models p rev iously w r it te n by the p re sen t author and by o th e r
experts w ith in th e f a c i l i t y in job model development a s w e ll as from
the Session M onitoring Form (see Appendix D) which was developed by
employees of the f a c i l i t y who were experts in programmed in s tru c tio n
and in s tru c t io n a l behavior m od ifica tion as w ell a s in sess io n m onitor
ing . The job models and the Session M onitoring Form both encompassed
behaviors s p e c if ic a l ly addressed by the BABS in th e p re sen t study.
The sc a le dimensions co n sis ted o f:
1. S® p re se n ta tio n req u irin g a ttend ing by the s tu d en t, a 1-second pause fo r s tu d en t a tten d in g by the in s tru c to r b e fo re p resen ting a c le a r and c o rre c t S as described w ith in the D is ta r le sso n .
2. Prompting re q u ir in g tim elines (not b e fo re 5 seconds and not a f te r 10 seconds o f no responding o r immediately a f t e r a wrong response) and achievement of a ta rg e t response as s t ip u la te d w ith in th e D is ta r le sso n .
3. Consequation in th re e modes o r dimensions w ith ap p rop ria te reinforcem ent w ith in 1 second o f a response, inc lud ing desc rip tiv e p ra is e and a tan g ib le reward, such as a token, ed ib le ,o r p h y sica l co n tac t.
I t should be noted th a t the sca le s developed fo r the p re sen t
study were much more s p e c if ic and lim ite d in scope than the BABS
developed in previous s tu d ie s fo r psychometric ev a lu a tio n and q u a rte r ly
to annually general a p p lic a tio n . A ctually , the sc a le s in th e p re sen t
study were no t intended fo r general performance ev a lu a tio n b u t were
intended: (a) to determ ine i f BABS p re sen ta tio n in a c o n tro lle d s e t
tin g could d i r e c t ly a f f e c t performance in the above ca teg o rie s which
were c ru c ia l to e f fe c t iv e teaching; and (b) to ev a lu a te th a t performance
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
fo r ln trad a ss ro o m management d ec is io n s . Because of th e s p e c if ic
content o f th e sc a le s and the value o f the anchors, th e p resen t BARS
may be even more o b je c tiv e than those seen in previous s tu d ie s so the
sca les may be considered B ehaviorally Anchored Rating S ca les . But
because of the lim ited scope o f the sca le s and the method of th e i r
a p p lic a tio n in the p re sen t study, the BARS presented h ere could not
su ff ic e as a to ta l BARS ev a lu a tio n system.
Experimental Design
A sim ple m u ltip le b a se lin e design across behaviors was u t i l iz e d
w ith Subjects A, B, C, and D. A re v e rsa l was a lso conducted w ith
Subject C. The o rders o f p re sen ta tio n o f sca le s fo r each su b jec t
a re p resen ted : S ubject A—consequation, prompting, and S® presen ta
tio n ; Subject B—consequation, S® p re sen ta tio n , and prompting; Sub
je c t C—consequation, prompting, and S® p re sen ta tio n ; and Subject D--
prompting, S® p re se n ta tio n , and consequation. The o rd er of in te rv en
tio n was determined by d e f ic i t s in the in s t ru c to r s ' re p e r to ire s and
re f le c te d an a ttem pt to provide the b e s t in s tru c tio n p o ss ib le fo r the
s tu d en ts .
Data C o llec tion
A standard sess io n m onitoring form was used to record da ta about
the in s t ru c to r s ' teach ing behav iors. As mentioned b e fo re , the observer
( in th is case , the experim enter and p resen t au thor) s a t behind the
s tuden ts in a p o s it io n fac ing the in s tru c to r . For each t r i a l which
began w ith each new in s tru c t io n , a p lus o r a minus was marked to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
in d ic a te whether o r n o t (a) the s tu d en t was a tten d in g ; (b) th e re was
a l-second pause by the in s tru c to r fo r s tu d en t a tten d in g ; (c) a cor
r e c t and c le a r S® was p resen ted ; (d) a c o rre c t response occurred and
(e) i f n o t, d id a prompt occur; (f) i f the prompt was n o t c o rre c t , was
i t l a t e o r did no ta rg e t besponse occur; and (g) was the consequation
co rrec t (ap p ro p ria te , immediate, d e sc r ip tiv e , and in c lu s iv e o f tan g ib le
rewards). A ppropriateness of consequation included nonreinforcem ent o f
in c o rre c t responses and reinforcem ent o f some form fo r c o rre c t
responses. In ap p ro p ria te reinforcem ent rendered consequation in c o rre c t
in a l l th re e dimensions fo r the re sp e c tiv e t r i a l .
BARS P re se n ta tio n
Subjects A and B were exposed to th e BARS immediately p r io r to
(w ith in 10 m inutes o f) the sess io n fo llow ing the one in which they
were observed and ra te d . This was done to a sse ss any antecedent
e ffe c ts o f the BARS upon s t a f f behav io rs. Subjects C and D were
exposed to the BARS immediately a f t e r (w ith in 10 seconds o f) the
session in which they were observed and ra te d . This was done to
assess any consequent e f fe c ts o f the BARS upon s t a f f behav io rs. A
summary o f experim enter-sub ject in te ra c t io n i s provided in Appendix E.
The BARS ra tin g s were always s ta te d in terms o f lower l im its s e t
down w ith in each s c a le . For example, i f 85% of re in fo rc e rs occurred
w ith in 1 second o f c o rre c t responses, th is was s ta te d as 70% as repre
sented by the lower l im i t on the s c a le . The num erical r a t in g always
represented the anchor d escrib ing the low est percentage v a lue fo r each
segment o f the dimension. S im ila r ly , i f S®'s were exact a t le s s than
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50% occurrance w hile a tten d in g , 1-second pauses, and c le a r S®'s
occurred a t 50%, the s c a le value o f 2 was c irc le d to rep resen t "wrong
SD"; and then the remaining d e sc r ip tiv e performance anchors were
underlined to a v a il the su b jec ts to the most s p e c if ic feedback possi
b le w ith th is form.
R e l ia b i l i ty
R e l ia b i l i ty of the observer was assessed ac ro ss sev e ra l exposures
to the same s e t of behav ioral events s in ce "assess in g the s t a b i l i t y
of observer judgments i s most e a s ily accomplished i f a constan t behav
io ra l event can be repeated , p resen ted , and observed" (Johnston &
Pennypacker, 1980, p. 195). Two videotaped sessio n s fo r each su b jec t
served as the repeated behav ioral even ts. The tap ings were i n i t i a l l y
made in the th ird sessio n s fo r Subjects B and D and in the fo u rth
sessions fo r Subjects A and C. This was done to allow the s tuden ts
and in s tru c to rs to become d esen s itiz ed to the new system before in tro
ducing a new and in tr ig u in g stim ulus and to allow the use o f the only
a v a ilab le videotape w ith a l l su b jec ts . The second tap ings were made
in Session 9 w ith Subjects B and D and in Session 10 w ith Subjects A
and C. This was done a t th i s p o in t simply to r e c a l ib r a te the observer
and to allow the use of the one tap e . The v ideo tape equipment was
placed to the r ig h t s id e of a l l o f the s tu d en ts to allow a view of
each o f them. R e l ia b il i ty was computed by d iv id in g a l l the agreements
by a l l the agreements p lu s disagreem ents fo r the sev e ra l observations
o f each in d iv id u a l sess io n .
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
RESULTS
The r e s u l ts o f the p re sen t study a re p resen ted g rap h ica lly in
F igures 1 - 4 . Three in te rv e n tio n elements a re d isp layed in each
m u ltip le b a se lin e graph. S® p re sen ta tio n i s d isp layed in terms of
s tuden t a tten d in g (rep resen ted by la rg e so lid do ts connected by so lid
l in e s ) , a 1-second pause fo r a tten d in g by th e in s tru c to r (rep resen ted
by open c i r c le s connected by dashed l in e s ) , and c o rre c t SD,s (rep re
sented by open squares connected by d o tted l i n e s ) .
The prompting element simply c o n s is ts o f the percentage of cor
r e c t prompts by each su b je c t. An arrow p o in ts down from the upper
l im it o f th is element in some cases to in d ic a te th a t one in c o rre c t
prompt occurred in le s s than 10 o p p o rtu n itie s to prompt. This i s
im portant to no te because th is a f fe c ts more g re a tly the v a r ia b i l i ty
in the percentages here than in the o th e r elem ents, s in ce in the l a t t e r ,
th e re were always a t l e a s t 10 o p p o rtu n itie s fo r responses, and in most
cases, 20 o r more. And, th e re was no c o n tro l o f o p p o rtu n itie s to
prompt, s in ce th is depended on in c o rre c t responses by the s tu d en ts .
Consequation i s d isp layed as percentage c o rre c t in th re e dimen
sions o f behavior (p ra is e , d e sc r ip tio n o f the behav ior, and a tan g ib le
reward) and in two dimensions (any two o f the previous dim ensions).
I t r e fe rs to reinforcem ent o f c o rre c t behaviors w ith nonreinforcem ent
o f c o rre c t behaviors o r reinforcem ent o f in c o rre c t behaviors counting
as in c o rre c t in a l l th ree dim ensions. In the graph fo r S ubject A, an
arrow p o in ts down from th e upper l im it o f consequation a t Session 10
to in d ic a te th a t the in s tru c to r changed the procedure to a VR2
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced
with perm
ission of the
copyright ow
ner. Further
reproduction prohibited
without
permission.
-O ' ,Q“o-
vtui«/)•o« mq...........
CL • •y».. •O’ -O-oc
ect* MIMNIAIIOMW M
U i
-O.-O-.CL. -O o-
INSTRUCTIONAl SESSIONS
Figure 1: Instructor Consequation, Prompting, and SD Presentation and Their E ffectsupon Student Responses as a Function o f Antecedent BARS Presentation forSubject A.
roto
Reproduced
with perm
ission of the
copyright ow
ner. Further
reproduction prohibited
without
permission.
M i l « l A N t a C l t l N !
■-cf
CONIIO*<tflOM
I/Ian
otot *•o ••UL
u- - a .
ItV M M t M l90*1 M l
INSTRUCTIONAL SE S S IO N S
Figure 2: Instructor Consequation, Prompting, and S Presentation andTheir E ffects upon Student Responses as a Function o f Antecedent BARS Presentation for Subject B.
Reproduced
with perm
ission of the
copyright ow
ner. Further
reproduction prohibited
without
permission.
-o-
-o'
uiM2i..O ..
....O■•o-
-Q-u
0£oc ••f l ••
O- .O-O•O' O' o-o o-
INSTRUCTIONAL SESSIONS
Figure 3: Instructor Consequation, Prompting, and SD Presentation and Their E ffectsupon Student Responses as a Function o f Consequent BARS Presentation forSubject C.
rocn
Reproduced
with perm
ission of the
copyright ow
ner. Further
reproduction prohibited
without
permission.
■O'•o<
V)
a.
>-
y i n
-Oo-
•o-
INSTRUCTIONAL SESSIONS
Figure 4: Ins truc to r Consequation, Prompting, and Presentation and Their Effects uponStudent Responses as a Function o f Antecedent BARS Presentation fo r Subject D.
I\
r
27
schedule o f token consequation to fade i t o u t as was customary in
many procedures in her classroom . Here the sca le on consequation
served as feedback to in d ic a te to the in s tru c to r w hether token o r
tan g ib le consequation was occu rring a t , above, o r below 50%. This
simply in d ic a te s th a t th e c r i te r io n was 50% in the th ree dimension
component, w hile the experim enter s t i l l sought p e rfe c tio n in the
remaining two dim ensions.
Student responses a re a lso d isp layed w ith each s u b je c t 's da ta
to dem onstrate any in d ir e c t e f fe c t o f the BARS on these through the
su b je c ts ' in s tru c t io n a l behav io rs. The percentage o f c o r re c t responses
i s d isp layed by open c i r c le s connected by dashed l in e s , and the per
centage of c o rre c t p lus co rrec ted responses ( to t a l percentage of
evoked responses) i s d isp layed by la rg e s o lid dots connected by s o lid
l in e s .
With Subject A, where BARS was used as an an teceden t, i t s pre
se n ta tio n apparen tly had an i n i t i a l l y d e le te r io u s e f fe c t upon her
consequation of s tuden t behav io rs. Consequation in b a se lin e ranged
from 77% to 84% in th ree dimensions and from 93% to 100% in two
dim ensions. With the advent o f an tecedent BARS p re se n ta tio n , conse
quation in th ree dimensions f e l l to 64% and in two dimensions to 82%.
Over the f iv e follow ing se ss io n s , these seem to recover somewhat.
The da ta begin to follow a p re d ic ta b le p a tte rn a t Session 10, where
the VR2 schedule w ith tokens was implemented. The VR2 token schedule
seems to have n eg ativ e ly a ffe c ted the o th e r two dimensions o f re in
forcement in Session , b u t reinforcem ent in two dimensions shows gradual
improvement th e re a f te r . Reinforcement in th ree dimensions a f te r
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28Session 10 stayed w ith in 10 percentage p o in ts o f c r i te r io n in f iv e
out o f e ig h t se ss io n s . In Session 17, S ubject A wanted to d iscon tinue
the use of tokens, so th e experim enter reminded h e r to use physica l
co n tac t in s te ad . Here consequation in th re e dimensions f e l l to 33%.
Reinforcement in th ree dimensions never dev ia ted more than 20 percen
tage p o in ts from the 50% c r i te r io n .
Percentage of c o r re c t prompting w ith S ubject A in the i n i t i a l
b a se lin e hovered around 50%, w ith an o v e ra ll in c rease in in te rv e n tio n
to about 80%. The graph sp ikes to 100% four times during the in te r
ven tio n , about 35 percentage p o in ts h igher than th e h ig h est percentage
during b a se lin e . In SD p re se n ta tio n , the experim enter w aited u n t i l
the d a ta showed some v a r ia b i l i ty b efo re in te rv en in g w ith the BARS
th e re , s in ce a l l the components o f SD p re se n ta tio n (a tten d in g ,
1-second pause, and c o rre c t S®) were hovering a t o r around c e i l in g .
There i s an upward trend in s tu d en t a tten d in g and in the 1-second
pause from 74% to 100% in the BARS in te rv e n tio n . However, due to the
end o f the school year, no fu r th e r d a ta were o b ta in ab le ; and given
more d a ta , perhaps a more d e f in i te trend would have appeared.
Student responses throughout th e study appear to improve in
terms o f c o rre c t responses and responses co rrec ted by prompts. The
l a t t e r n e c e s sa r ily co-varied to some ex ten t w ith the percentage o f
c o rre c t prompts by the in s t ru c to r . But the percentage o f co rre c t
s tu d en t responses a lso appears to co-vary w ith the percentage o f cor
r e c t prompts by Subject A. This i s p a r t ic u la r ly ev iden t in Sessions 9
through 14, where in every in s tan ce except Session 12, the percentage
o f c o r re c t s tuden t responses shows a d ire c t re la tio n s h ip w ith the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
percentage o f c o rre c t prompts by the in s tru c to r . C orrect s tu d en t
responses in c rease from 50% in Session 9 to 73% and 90% in Sessions 10
and 11 re sp ec tiv e ly w ith the advent o f a VR2 schedule o f token re in
forcement. But a f t e r Session 11, s tu d en t response r a te s f e l l to a
lower b u t more c o n s is te n t r a te o f about 65% fo r the rem ainder o f the
study.
Consequation o f s tu d en t responses by Subject B shows a substan
t i a l amount o f v a r ia b i l i ty in b a se lin e . The range of percentage
p o in ts i s from 20 to 87 in th ree dimensions o f consequation and 68
to 100 in two dimensions o f consequation w ith no apparent general
tren d . With the in tro d u c tio n of an tecedent BARS p re se n ta tio n , th is
v a r ia b i l i ty i s reduced w ith a su b s ta n tia l in c rease in c o rre c t conse
quation in two and th ree dim ensions. The range of scores in BARS
in te rv e n tio n i s 60% to 100% in th re e dimensions o f consequation and
77% to 100% in two dimensions o f consequation.
SD p re sen ta tio n in a l l th ree ca teg o rie s (s tuden t a tten d in g ,
1-second pause, and c o rre c t S^) w ith Subject B shows some i n i t i a l
v a r ia b i l i ty but then approaches c e i l in g w ith some v a r ia b i l i ty w hile
s t i l l in b a se lin e . In te rv e n tio n was req u ired a t Session 11 because
o f the nearing o f the c lo se o f the school y ear. However, even w ith a
lack of room fo r improvement, the data do show some in c rease and
narrowing o f v a r ia b i l i ty in percentage p o in ts fo r a ttend ing and
c o rre c t S ^ 's and fo r the 1-second pause except fo r Session 13.
The b a se lin e da ta fo r S ubject B in prompting may be deceiv ing .
In Sessions 5, 7, 8, and 11, only one in c o rre c t prompt occurred .
Because o f th i s , the d a ta could have been very much d if f e r e n t given
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
more o p p o rtu n itie s to prompt. Conceivably, le s s v a r ia b i l i ty could
have occurred i f only a few more c o rre c t prompts were d e liv e red in
Sessions 5 and 8 and even in Session 6, where only two o p p o rtu n itie s
fo r prompting occurred bu t which y ie ld ed 0%. On the o th e r hand,
given more in c o rre c t prompts in Sessions 7 and 11, the da ta again may
have looked very d i f f e r e n t . T herefore, due to the i n s ta b i l i ty and
v a r ia b i l i ty in the b a se lin e prompting d a ta , the experim enter w aited
as long as p o ss ib le to implement an tecedent BARS p re sen ta tio n fo r
in s tru c to r prompting. Upon BARS in te rv e n tio n , no narrowing o f v a r ia
b i l i t y from the l a s t fou r b a se lin e d a ta days i s ev iden t.
The r a te o f c o rre c t s tu d en t responses shows l i t t l e re la tio n sh ip
w ith the percentage o f c o rre c t prompts by S ubject B. However, i t
does show a d i r e c t re la tio n s h ip w ith s tu d en t a tten d in g in Sessions 1
through 8 and some c o -v a ria tio n in the remainder of the se ss io n s . In
a d d itio n , c o rre c t s tuden t responses co-vary w ith the 1-second pause
throughout the study except in Sessions 1 , 6, 10, and 13. And they
co-vary w ith consequation in th ree dimensions in Sessions 2 through 9
and in Session 14.
With Subject C, one f a c t i s re a d ily apparen t. That i s , th a t no
re v e rsa l occurred in any elem ent. And in every elem ent, the same
general statem ent can be made about the e f fe c ts of consequent BARS
p re se n ta tio n . There was moderate v a r ia b i l i ty in b a se lin e w ith some
apparent improvement in In s tru c to r responses given the in tro d u c tio n
o f the BARS. There i s s t i l l fu r th e r improvement in the re v e rsa l and
in BARS re in tro d u c tio n re sp e c tiv e ly . Percentage o f c o r re c t s tu d en t
responses does no t apparen tly co-vary w ith any element o r any segment
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
th e reo f. And th e re Is no general Increase over time in the percentage
of c o rre c t s tu d en t responses.
The b a se lin e prompting da ta fo r S ubject D show a downward trend
from about 80% to about 40%. With the in tro d u c tio n of consequent BARS
p re se n ta tio n , a su b s ta n tia l in c rease in percentage of c o rre c t prompts
i s observed from Session 5 to 7, w ith a gradual decrease from Session 7
to 10. Again, a s u b s ta n tia l in c rease from 50% to 100% occurs in the
remaining four se ss io n s .
SD p re sen ta tio n by Subject D shows a moderate amount o f v a r ia b i l i ty
in b a se lin e , e sp e c ia lly fo r a tten d in g and the 1-second pause. With
the in tro d u c tio n o f the BARS, th e re i s both a red u c tio n in v a r ia b i l i ty
o f and a general in c rease in the percentage of c o rre c t SD,s , s tu d en t
a tten d in g , and the 1-second pause. The percentage of c o rre c t conse
quation in two and th re e dimensions in b a se lin e i s very high and shows
l i t t l e v a r ia b i l i ty . With the in tro d u c tio n o f BARS p re se n ta tio n , how
ever, consequation in th ree dimensions f a l l s f r a s t i c a l l y , w hile conse
quation in th ree dimensions remains very h ig h . C orrected studen t
responses appear to co-vary w ith consequation in two dim ensions, and
th e re appears to be a s l ig h t Improvement in both c a teg o rie s of s tuden t
responses over the course o f the study.
The r e l i a b i l i t y d a ta fo r the observer were c a lc u la te d as the
to ta l number of agreements over the to ta l number of agreements p lus
disagreem ents in th ree repeated observations and d a ta record ings o f
two d if f e r e n t s e ts o f s t im u li (two d i f f e r e n t v ideo tapes o f a c tu a l
experim ental sessions fo r each su b jec t) between su b je c ts over a l l the
dimensions. The r e l i a b i l i t y d a ta f o r Viewing A a re : S ubjec t A, 92%;
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
Subject B, 98%; Subject C, 94%; and Subject D, 89%. The r e l i a b i l i t y
d a ta fo r viewing B a re : Subject A, 95%; S ubject B, 91%; S ubject C,
97%; and Subject D, 90%.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DISCUSSION
I t i s apparent a f t e r a genera l overview of the r e s u l ts of th is
study th a t the in tro d u c tio n o f the BARS had a s l ig h t fu n c tio n a l re la
tio n sh ip w ith an in c rease in the percentage o f c o rre c t in s tru c to r
responses. This i s e sp e c ia lly tru e in the case of Subject A, where
consequation in th ree dimensions conforms c lo se to the 50% c r i te r io n
l in e when nowhere e a r l i e r in her case d id th is occur. BARS e f f e c t on
in s tru c to r responses was a lso apparent w ith S ub jec t B in h er consequa
tio n of s tu d en t behaviors and to a le s s e r ex ten t in her SD p resen ta
tio n . With Subject C, the BARS appear to have had a s l ig h t e f f e c t
upon in s tru c to r responses in a l l the elem ents; b u t the continued
gradual in c rease in these responses during re v e rsa l i s puzzling . And
f in a l ly , except fo r consequation, Subject D’ s in s tru c tio n a l responses
appear to improve w ith the in tro d u c tio n o f BARS p re se n ta tio n .
Before an an a ly s is o f the success o f BARS i s made, a d iscu ssio n
o f i t s shortcomings i s in o rd e r. The f i r s t major apparent f a i lu r e of
BARS was in i t s e f fe c ts upon consequation fo r Subject D. The v a riab le
th a t most d ram atically occurs along w ith the dip in consequation i s a
break in sessio n s o f 5 d a ta days between Sessions 9 and 10 due to a
long Memorial Day c lo sin g of the f a c i l i t y where th is study was con
ducted. Another v a r ia b le th a t occurred here was th a t S ubject D was
tra n s fe rre d to a d if f e r e n t classroom a f te r th a t break than the one
she worked in before i t . Due to th is t r a n s fe r , h e r schedule c o n flic te d
w ith the schedule normally m aintained fo r th is study; and she engaged
in a g re a t deal of behavior to enable h e r s e lf to p a r t ic ip a te in th is
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
study . These v a ria b le s may have had some e f f e c t on her performance in
the in s tru c tio n a l se ss io n s . Whether the cause o f the drop in perform
ance was due to these o r o th e r unsuspected v a r ia b le s , i t apparen tly
a lso had a negative e f f e c t on prompting by Subject D, although i t was
no t such a profound e f f e c t . The p o s s ib i l i ty a lso e x is ts th a t the
observer was le n ie n t o r la x in the record ing o f consequation in th ree
dimensions during b a se lin e and then became more observant o f re in fo rc e
ment w ith the In tro d u ctio n o f BARS fo r in s t ru c to r consequation. But
th is i s no t c o n s is te n t w ith the d a ta fo r the o th e r su b je c ts , nor w ith
the r e l i a b i l i t y d a ta .
Another problem w ith the consistency of BARS e f fe c ts on in s tru c to r
behaviors i s the lack of a re v e rsa l in any element fo r S ubject C.
There a re sev e ra l p o ssib le exp lana tions. F i r s t , the BARS may have had
l i t t l e o r no e f f e c t on h e r behavior w ith the improvement in her responses
being due to simple p ra c tic e e f f e c ts . Conceivably, reinforcem ent fo r
her responses could have been programming by growing fa m ilia r i ty w ith
the m a te ria ls or by success w ith s tuden t compliance as a fu n c tio n o f
follow ing the w r it te n D is ta r m a te r ia ls . Another p o s s ib i l i ty i s th a t
the feedback contained in the BARS evoked rule-governed behavior and
th a t th is p e rs is te d from the in tro d u c tio n of BARS. This l a t t e r explana
tio n seems the more parsim onious s in ce th e i n i t i a l in tro d u c tio n of
BARS in consequation and in S® p re se n ta tio n very quickly reduced
v a r ia b i l i ty in the d a ta and improved the in s t r u c to r 's responses. This
i s a lso tru e to a le s s e r ex ten t w ith prompting. In o th e r words, the
m u ltip le b a se lin e appears to have shown an e f fe c t o f BARS in tro d u c tio n .
The major e f fe c t o f BARS in th is study seems to have been due to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35i t s feedback c h a r a c te r is t ic s . The sca le s d id no t con ta in p a r t ic u la r ly
p o s itiv e o r negative la b e ls to serve as re in fo rc e rs o r pun ishers. The
su b jec ts d id , however, vo ice d isp leasu re w ith low rankings and p leasu re
w ith high rankings. The observer who presen ted the sca le s would, upon
the achievement o f a 10 (the h ig h est ra tin g ) , show the sc a le o r sca le s
to the su b jec t and announce, "You're a 1 0 ." The su b jec ts a l l seemed
to enjoy th i s . But o th e r than th i s , th e re was no in h eren t re in fo rc in g
o r punishing stim ulus a sso c ia ted w ith the s c a le s . The su b jec ts a l l
vo lunteered the inform ation th a t the sc a le s served two fu n c tions fo r
them. F i r s t , they showed the in s tru c to r ex ac tly what she d id in
conjunction w ith a d iscu ss io n of her performance w ith the observer.
Secondly, they showed h e r e x ac tly what i t was she had to do to a t t a in
the h ig h est ranking. F in a lly , th e observer was acquainted so c ia lly
w ith a l l the su b jec ts . This may have rendered low scores s o c ia lly
punishing and high sco res s o c ia l ly re in fo rc in g .
C onsisten t w ith the fin d ings of Krumhus and M alott (1980), th e re
was no s ig n if ic a n t d iffe re n c e between the e f fe c ts of BARS p re se n ta tio n
as an an tecedent and as a consequence. However, the d a ta fo r an tece
dent BARS p re sen ta tio n follow a gen era lly more p re d ic ta b le p a tte rn
than the da ta fo r consequent BARS p re sen ta tio n . I f the e f fe c ts of BARS
a re due mainly to th e i r feedback c h a r a c te r is t ic s , then th a t feedback
may be more e f fe c tiv e in evoking rule-governed behav ior when i t i s
presen ted immediately p r io r to the ta rg e ted behav io rs, w hile i t s
e f fe c ts in evoking ru le-governed behavior may dim inish over tim e.
However, th is is an is su e th a t should be d e lib e ra ted in the feedback
l i t e r a tu r e . The d a ta in the p resen t study provide nothing but
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36sp ecu la tio n as to the d iffe ren c es between an tecedent and consequent
BARS p re sen ta tio n .
The v a lid a tio n fo r the study o f the e f fe c ts o f using BARS as a
tra in in g to o l i s ev iden t in the da ta fo r s tu d en t responses. The f i r s t
question personnel managers might have about BARS i s "W ill i t improve
performance?" That has been shown to have occurred to some ex ten t.
The second question i s "W ill i t have a p o s it iv e e f f e c t on the rece iv in g
system, the environment, o r the c l ie n te l le ? " In o ther words, w i l l i t
r e s u l t in re a l o rg an iza tio n a l improvement and b e n e f its? P o s itiv e
e f fe c ts in the rece iv in g system a re demonstrated to a moderate bu t
f a i r l y c o n s is te n t e x ten t w ith the improvement in s tu d en t responses.
In D is ta r Math and Language programs, an o v e ra ll success r a te o f 75%
c o rre c t s tuden t responses i s the goal. In the case of S ubjects A and
D, co rrec ted s tuden t responses went from under 75% to 100%; and they
went from under 90% to 100% fo r Subjects B and C. A 25% e r ro r r a te
i s supposed to allow fo r a s l ig h t amount o f programmed f a i lu r e to give
the s tu den ts some inform ation as to what no t to do and to make re in
fo rc e rs more e f fe c t iv e . But in th is study , the high response ra te s
served to v a lid a te the use o f BARS as a tra in in g to o l.
The p re sen t study d i f f e r s experim entally from previous s tu d ie s
in BARS ap p lic a tio n s in ce those, alm ost ex c lu s iv e ly , have d e a l t w ith
psychometric comparisons o f BARS w ith o th e r ra tin g form ats and have
no t d e a lt w ith e f fe c ts o f BARS on behavior o r o th e r o rg an iza tio n a l
r e s u l t s . Ivancevich (1980b) rep o rted an improvement in scheduling
performance by p ro fess io n a l employees, bu t th is was no t q u a n tif ie d .
This study was an a ttem pt on a b a sic and em pirica l le v e l to determ ine
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
whether BARS would have any tru e e f f e c t upon performance. The impor
tance o f th is study may no t l i e in the re s u l ts o f the experiment (or
the lack th e re o f) , b u t i t may w ell l i e in the precedent i t has s e t in
seeking and e s ta b lish in g beh av io ra l goals fo r and e f fe c ts o f BARS
a p p lic a tio n .
Landy and Barnes (1979) have s e t down sp e c if ic p o in ts of lo g ic in
using BARS. The people who use the sca le s have to some ex ten t in most
cases been i t s developers as w e ll. More inform ation i s provided w ith
behav io ral anchors than w ith sim ple heading d e sc r ip to rs . These anchors
have s im ila r meanings to a l l r a t e r s , and the performance dimensions in
BARS a re e a s ily d is tin g u ish ab le from one another.
And a r is in g need fo r accu ra te performance a p p ra isa ls makes BARS
perhaps the b e s t choice o f sev e ra l e v i l s s in ce , as C hurch ill sa id of
democracy, " I t ' s a t e r r i b l e system b u t i t ' s the b e s t th e re i s . "
Keaveny and McGann (1975) p o in t ou t th a t accu ra te performance a p p ra isa ls
allow fo r proper reward and promotion fo r e f fe c t iv e and high perform ing
personnel. This would be a break fo r employers who a re co n stan tly
seeking b e t te r q u a lif ie d persons to f i l l im portant p o s itio n s . On the
o th e r end o f the spectrum, w hite c o l la r employees a re d isp u tin g d is
charges more freq u en tly ; and an a ccu ra te a p p ra isa l of poor performance
would f a i r ly support such d isch arg es . So in terms of r e a l organiza
tio n a l c o s ts , accu ra te and e f fe c t iv e ap p ra isa ls o f performance a re a
grave n ecess ity .
The v a r ia t io n o f the sc a le s used in the p re sen t study from the
ty p ic a l BARS approaches a s im ila r i ty to the Behavior O bservation
S cale, o r BOS. The in te n t o f the BOS i s to equalize the value o f
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38re sp e c tiv e le v e ls of performance across behavior item s (Latham, Fay, &
S a a ri, 1979). The BOS c o n s is ts o f performance dimensions named
" c r i t e r i a . " w ith subsets of "behav iora l item s." The c r i t e r i a and
item s a re generated in a manner s im ila r to the development o f per
formance dimensions and b ehav io ra l anchors fo r a BABS. The b ehav io ra l
item s in a BOS number from 5 to 10 in each of 5 to 10 c r i t e r i a . The
item s a re underscored by groups o f o b serv a tio n a l frequencies ( i . e . , 0 -
20%, 21 - 40%, 41 - 60%, 61 - 80%, and 81 - 100% frequency o f observa
tio n o f the behav io ral item ), the ap p ro p ria te category of which i s
checked a t the time of the performance ev a lu a tio n (Kane & B ernardin ,
1982).
As w ith every performance ev alu a tio n to o l reviewed by the pre
sen t au tho r, problems w ith the BOS have been p resen ted . Kane and
B ernardin (1982) claim th a t th e goal of the BOS, which i s to equalize
re sp ec tiv e le v e ls o f frequency, i s f a l l i b l e in th a t eq u a lity o f f r e
quency o f observation does no t n e c e s s ita te e q u a lity o f the w eight of
the frequency le v e ls between item s. For In s tan ce , they p o in t out th a t
equal frequencies o f o b ta in in g homicide a r r e s t w arrants and being
c lea red by the In te rn a l Review Board fo r the use of f a t a l fo rce by a
p a r t ic u la r p o lic e o f f ic e r i s unacceptable. E ighty percen t frequency
of o b ta in in g homicide a r r e s t w arrants may be su p e r io r , but being
c leared by the In te rv a l Review Board 80% of the time i s horrendous.
Latham, S aa ri, and Fay (1980) claim th a t th is e q u a li ty of frequency
le v e ls does e x is t but no d e f in i t iv e re so lu tio n s have been achieved
and the debate has become heated (Bernardin & Kane, 1980; Latham,
S a a ri, & Fay, 1980).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39In review , the BARS system c o n s is ts of performance dimensions
underscored by a sca le which re p re sen ts poor to e x ce lle n t performance
w ith b ehav io ra l anchors a ttach ed to the sca le values d esc rib in g
examples of poor to ex ce lle n t perform ance. The sca le s used in the
p resen t study c o n s is t of performance dimensions underscored by sca le s
rep resen tin g poor to e x ce lle n t performance w ith behav io ral anchors
c o n sis tin g of frequencies of observation of s e ts of behaviors a ttached
to the sca le values to d esc rib e poor to e x c e lle n t performance. I t i s
these frequencies of performance which resemble the frequencies of
observation o f behaviors in the BOS. While these frequencies do
resem ble, and indeed serve a s im ila r purpose a s , th e BOS, th e re a re
im portant d iffe ren c es to make c le a r . F i r s t , the frequencies provided
w ith in the sca le s in the p resen t study a re a departu re from a ty p ic a l
BARS which c o n s is t o f a d je c t iv a l ra th e r than num erical d e sc rip tio n s
of poor to e x ce lle n t performance comprising a performance dimension.
This d epartu re was u t i l i z e d to maximize the o b je c tiv e observation and
ev alu a tio n of su b jec t behavior and was conceived to make an em pirical
assessment o f the BARS as a t ra in in g to o l ra th e r than to c re a te a new
performance ev alu a tio n to o l . Second, the re sp ec tiv e frequency le v e ls
o f performance a re no t and were not intended to be p a r a l le l across
performance dimensions; and the ob serv a tio n a l frequency le v e ls must
be p a ra l le l across behavior items to q u a lify as a BOS.
D espite the fa c t th a t the p resen t study concentrated on assessin g
the BARS as a tra in in g to o l , i t does not mean th a t i t may not provide
im portant inform ation ap p licab le to the BOS and o th e r ev a lu a tio n to o ls .
As has been pointed out e a r l i e r , i t may w ell be the case th a t the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
method o f o b ta in ing the inform ation p e r tin e n t to performance ev a lu a tio n ,
the inform ation a c tu a lly used, and the way i t i s used ( i . e . , w i l l any
change in the environment occur as a r e s u l t of the evalua tion ) a re
more im portant than the name of the to o l used or the way the informa
t io n i s d isp layed .
P eterson (1982) provides some suggestions fo r the use of term in
ology in feedback s tu d ie s . He claim s th a t the term "feedback" has
become "p ro fess io n a l s lang" and th a t i t should not be used as a des
c r ip t io n o f the in te rv e n tio n a l procedure. He suggests th a t au thors in
o rg an iza tio n a l behavior management l i t e r a tu r e do not s u f f ic ie n t ly
d e fine the s p e c if ic feedback procedures used and th a t these sp e c if ic
d e sc rip tio n s must be provided to in te rp re t and analyze the procedures
and th e i r r e s u l t s . T h is , he says, w i l l f a c i l i t a t e the a n a ly s is of the
e f fe c ts of feedback and th a t th is i s o f paramount importance in o rder
to move away from unfounded assum ptions in the l i t e r a tu r e about the
e f fe c ts feedback has on performance so th a t feedback e f fe c ts can be
assessed . When d e sc rip tio n s of s p e c if ic feedback procedures a re not
provided in the l i t e r a t u r e , the behav io ra l e f fe c t of feedback ( i . e . ,
re in forcem ent, punishment, stim ulus c o n tro l) cannot be assumed; and the
resea rch cannot be in te rp re te d in any meaningful way. S pecifiy ing
experim ental procedures and th e i r r e s u l t s should be, in th is a u th o r 's
op in ion , the focus o f subsequent s tu d ie s in using feedback to a l t e r
behavior and improve perform ance. These p o in ts can a lso be applied
to s tu d ie s involving ra t in g sc a le evalu a tio n s o f performance since
adequate d e sc rip tio n s o f the way in which the ra t in g sc a le s a re used
has freq u en tly not been sp e c if ie d in performance ev alu a tio n s tu d ie s .
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41Remedies to these problems w ith feedback and ev alu a tio n s tu d ie s were
sought in the p resen t study by providing the sca le s used and d e ta ile d
d e sc rip tio n s o f the procedures.
More s tu d ie s l ik e the p re sen t one should be conducted in the
fu tu re to determ ine the ex ten t o f the e f f e c t BARS can have on an
o rg an iza tio n . The au thor would no te th a t i t has been encouraging to
see so many s tu d ie s devoted to the development of d e a r and o b jec tiv e
statem ents about o rg a n iz a tio n a l perform ance. As Drucker (1973) has
sa id :
An employer has no business w ith a man's p e rso n a lity . Employment i s a s p e c if ic c o n tra c t c a l l in g fo r s p e c if ic perform ance, and fo r nothing e ls e . Any a ttem pt o f an employer to go beyond th is i s u su rpa tion . I t i s immoral as w e ll a s i l l e g a l in tru s io n in to p rivacy . I t i s abuse of power. An employee owes no " lo y a lty "— he owes performance and nothing e ls e . . . . Management and manager development should concern themselves w ith changes in behav* io r l ik e ly to make a man more e f fe c t iv e , (p . 424-425)
I t i s the conclusion o f the au thor th a t the BARS may serve as an
e f fe c t iv e tra in in g to o l. .This may be accomplished by using the sca le
i t s e l f as a feedback mechanism as was done in th is study , o r by using
the inform ation gained in t r a d i t io n a l BARS development procedures to
develop a programmed in s tru c t io n o r o th e r behavior-based tra in in g pro
gram. But before th is can be a c e r ta in ty , more evalu a tio n s of BARS
as a tra in in g to o l must be made. The au thor suggests a move away from
psychometric evalua tions o f BARS un less new developments a re m anifest
in BARS and o th e r performance ev alu a tio n procedures. As B em ardin and
Kane (1980) p o in t o u t, " I t i s our view th a t another round of em pirica l com
parisons between form ats th a t do no t d ep art appreciab ly from prev iously
te s ted methods i s unwarranted" (p . 810). A behav io ral ev a lu a tio n of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
BARS by experienced developers would be paramount. With the advent of
BARS th a t a re developed through the more t r a d i t io n a l means y ie ld in g
psychom etrically sound s c a le s , more profound tra in in g e f fe c ts o f BARS
than in th is study may become apparent.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX A: SD PRESENTATION
90%—A ttending 90%—1 -sec . pause 90%~SD, s ig n a l exact 90%—C lear p re sen ta tio n
i—lO
9 -
90%—A ttending 90%—1 -sec . pause 70%—S°, s ig n a l exact 70%—C lear p re sen ta tio n
- 8
7 -
70%—A ttending 70%—1 -sec . pause 90%—SD, s ig n a l exact 90%—C lear p re sen ta tio n
- 6
50%—A ttending 50%—1 -sec . pause 50%—Unclear SD 50%—S®, s ig n a l inexact
3 -
Hrong S®
1 -
Wrong S®No a ttend ing
90%—A ttending 90%—SD, s ig n a l exact 90%—C lear p re sen ta tio n 70%—1 -sec . pause
90%—A ttending 70%—1 -sec . pause 70%—SD, s ig n a l exact 70%—C lear p re sen ta tio n
70%—A ttending 50%—1 -se c . pause 50%—SD, s ig n a l inexac t 50%—SD un c lea r
- 4
No S1
- 2
No a tten d in g
1— 0
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX B: PROMPTING
90%—A fte r 3 sec . of no responding
90%---A fter wrong response 90%—Target response achieved
9 -
70%—A fte r 3 sec . of no responding
70%—A fte r wrong response 90%—Target response achieved
7 -
50%-—Sometime a f t e r 3 sec . o f no responding
50%~More than 3 sec. a f t e r wrong response
70%—Target response achieved5 -
50%—-Sometime a f te r 3 sec . of no responding
50%—More than 3 sec. a f t e r wrong response
50%—-Target response achieved3 -
70%—No ta rg e t response 50%—In c o rre c t prompt
1 -
50%—No prompt a f te r no o r wrong response
r lO
90%—A fte r 3 sec . o f no responding 90%—A fte r wrong response .70%—T arget response achieved
- 8
70%-—A fte r 3 sec . o f no responding 70%—A fte r wrong response 70%—Target response achieved
- 6
-4
50%—Before 3 sec . o f no responding 50%—Before wrong response
- 2
50%—A fte r c o rre c t response
u 0
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX C: CONSEQUATION
90%~Sr+ w /in 1 sec. o f response 90%—Tangible Sr +, d e sc r ip tio n ,
inm ediate p ra is e ( th ree dimensions)
9 -
70%—Sr+ w /in 1 sec. o f response 90%—-Sr+ in th ree dimensions
50%—Sr + up to 5 sec. a f te r response
90%—Sr + in th ree dimensions
a f te r50%—Sr + up to 5 sec response
90%—Sr + in one dimension
30%—Sr + fo r wrong response
50%—Sr+ fo r wrong response
7 -
5 -
3 -
1 -
r 10
90%—Sr+ w /in 1 sec . o f response 90%—Sr+ in two dimensions
- 8
70%—Sr+ w /in 1 sec . o f response 90%—Sr in two dimensions
- 6
50%—Sr+ up to 5 sec . a f t e r response 90%—Sr + in two dimensions
- 4
30%—No Sr + fo r response
- 2
50%—no Sr+ fo r response
- 0
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX D: SESSION MONITORING FORM
Tutor:___________________ __________ _ P u p il:Date: Classroom:Session Time: ______________ ____________ Monitor:
Monitor da te:
I s the procedure a f fe c tin g behavior change in the d esired d ire c tio n ? YES NO
Number o f days on th is s te p / phase:
General comments:
Tutor s ig n a tu re :
M onitor da te:
I s the procedure a ffe c t in g behavior change in the d e sired d irec tio n ? YES NO
Number o f days on th is s te p / phase: _____
General comments:
Tutor s ig n a tu re :
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TRIA
L/IN
TERV
AL
PRETASK T PROMPTS POSTTASK
INAP
P RE
SP
CONS
EQUE
NCE
A.
Atte
ndin
g?
B. 1
-sec
ond
paus
e?C.
SD
corr
ect?
D.
R co
rrec
t?E.
Prom
pt
corr
ect?
1. L
ate
2. No
targ
et
R
F.
Con
seq.
co
rrec
t?
1. In
appr
opri
ate
2. No
t Im
med
iate
3. No
des
crip
tio
i
123456789
10
TRIA
L/IN
TERV
AL
PRETASK T PROMPTS POSTTASK
INAP
P RE
SP
CONS
EQUE
NCE
A.
Atte
ndin
g?B.
1-
seco
nd
paus
e?C.
SD
corr
ect?
D.
R co
rrec
t?E.
Prom
pt
corr
ect?
1. L
ate
2. No
targ
et
R
F.
Con
seq.
co
rrec
t?
1. In
appr
opri
ate
2. No
t im
med
iate
3. No
desc
ript
ion
123456789
10
APPENDIX E: INTERACTIONS BETWEEN OBSERVER AND SUBJECT
When the BARS were p resen ted as an teceden ts and as consequences,
the experim enter in te ra c te d w ith the su b jec ts to exp lain exactly what
the s c a le feedback rep resen ted . Examples of th is in te ra c t io n follow .
The experim enter in te ra c te d w ith each su b je c t in d iv id u a lly .
F i r s t , he would p re sen t the s c a le (s ) and sim ply r e c i te the inform ation
contained in the s c a le . Then, he would r e la te th a t inform ation to the
previous d ay 's performance on each s c a le . In o th e r words, "You scored
a 4 today, and th a t i s worse than y este rd ay ," o r "You scored a 9 today,
and th a t i s b e t te r than y e s te rd a y ." This was done to give the sub
je c t s some id ea o f the p rogress o f th e i r performance. A fte r these
d a ily in te ra c t io n s , the experim enter asked the su b jec ts i f th e re were
any questions about the s c a le , th e i r perform ance, o r the experim ent.
Almost ex c lu s iv e ly , they would ask fo r an a n a ly s is of th e i r performance,
such a s , "How should I p re sen t th is new task?" o r "What should I do
about a c h i ld 's behavior problem?" Or they would ask what they should
do s p e c if ic a l ly to improve th e i r r a t in g . O ccasionally , they would make
suggestions fo r amendment o f the w r it te n procedure. And every su b jec t
asked a t l e a s t once how the experiment in general was going, but th is
was ju s t out o f c u r io s i ty about how an experiment i s run.
Regarding questions asked about how to deal w ith the s tu d en ts ,
because the s tu d e n t 's b e s t in te r e s t was o f g re a te s t im portance, the
observer would do whatever was necessary to help the su b jec t deal w ith
the s tu d en t, e sp e c ia lly in the case of the behavior problem. I f the
question d e a lt w ith the s p e c if ic D is ta r procedures involved, the sub
je c t was f i r s t d ire c ted to read the m anual's in s tru c tio n s ; and nex t,
i f necessary , the su b jec t was d ire c ted to read over the ra tin g sca le s
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
s in ce these included a l l the inform ation necessary to conducting an
In s tru c tio n a l sessio n w ith in the f a c i l i t y . And regard ing questions
about the resea rch p ro je c t , they were answered as honestly as p o ssib le
s in ce th is was a p o licy o f the experim enter.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
BIBLIOGRAPHY
A llan , P . , & Rosenberg, S. The development of a ta sk o r ie n te d approach to performance ev a lu a tio n in the c i ty o f New York. P ublic Personnel Management, 1978, 1 , 26-32.
A tkin, R. S . , & Conlon, E. J . B ehaviorally anchored ra tin g s c a le s :Some th e o re tic a l is su e s . Academy of Management Review, 1978, 3., 119-128.
B ernardin, H. J . B ehavioral ex p ec ta tion s c a le s : A f a i r e r comparison.Journal of Applied Psychology, 1977, 62, 422-427.
B ernardin, H. J . , A lvarez, K. M., & Cranny, C. J . A recomparison of behav ioral ex p ec ta tion sca le s to summated s c a le s . Jou rnal o f Applied Psychology, 1976, 61, 564-570.
B ernardin, H. J . , & Kane, J . S. A second look a t b ehav io ra l observatio n s c a le s . Personnel Psychology, 1980, 33, 809-814.
B ernardin, H. J . , & Smith, P. C. A c la r i f i c a t io n o f some is su e s regarding the development and use o f b eh av io ra lly anchored ra t in g sca le s (BARS). Journal o f Applied Psychology, 1981, 66, 458-463.
Blood, M. R. S p in -o ffs from b ehav io ra l ex p ec ta tion s c a le procedures.Journal o f Applied Psychology, 1974, 59, 513-515.
Borman, W. C. Exploring upper l im i ts o f r e l i a b i l i t y and v a l id i ty in job performance r a t in g . Jou rna l o f Applied Psychology, 1978, 63, 135-144.
Borman, W. C ., & Dunnette, M. D. Behavior based versus t r a i t o rien ted performance ra tin g s : An em pirica l study . Jou rna l o f AppliedPsychology, 1975, 60, 561-565.
Borman, W. C ., & V allon, W. R. A view o f what can happen when behavio r a l expec ta tion sc a le s a re developed in one s e t t in g and used in ano ther. Journal of Applied Psychology, 1974, 50, 197-201.
Campbell, J . P . , Dunnette, M. D ., Arvey, R. D ., & H e lle rv ik , C. V.The development and ev a lu a tio n o f beh av io ra lly based ra tin g sca le s . Journal o f Applied Psychology, 1973, 57, 15-22.
C araine, C. W., & Fink, W. T. Increasing the r a te o f p re sen ta tio n and use o f s ig n a ls in elem entary classroom teach e rs . Journal o f Applied Behavior A nalysis, 1978, 11, 35-46.
Das, H ., F ro s t, P. J . , & Barnowe, T. B ehaviorally anchored ra t in g sca le s fo r assessing behav io ral sc ience teach ing . Canadian Journal o f Behavioural Science, 1979, 11, 79-88.
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
D eC otiis, T. A. An a n a ly s is o f the ex te rn a l v a l id i ty and applied re levance of th ree r a t in g form ats. O rgan izational Behavior and Human Performance. 1977. 19, 247-266.
D eC otiis, T. A. A c r i t iq u e and suggested re v is io n o f beh av io ra lly anchored ra t in g sc a le s development procedures. Educational and Psychological Measurement, 1978, 38, 681-690.
D ickinscn, T. L . , & Z e llin g e r , P. M. A comparison o f the b eh av io ra lly anchored ra t in g and mixed standard sc a le form ats. Journal of Applied Psychology, 1980, 65, 147-154.
Drucker, P. Management: Tasks, r e s p o n s ib i l i t ie s , and p ra c t ic e s .New York: Harper & Row, 1973, 424-425.
Fabry, P. L . , & Reid, D. H. Teaching fo s te r grandparents to t r a insevere ly handicapped persons. Journal o f Applied Behavior A nalysis,1978, 11, 111-123.
Flanagan, J . C. C r i t ic a l requirem ents: A new approach to employee ev a lu a tio n . Personnel Psychology, 1949, 2 , 419-425.
Ivancevich, J . M. B ehavioral exp ec ta tio n sca les versus nonanchored and t r a i t ra tin g system s: A s a le s a p p lic a tio n . Applied Psycholog ic a l Measurement, 1980a, _4, 131-133.
Ivancevich , J . M. A lo n g itu d in a l study o f b eh av io ra l expec ta tionsc a le s : A ttitu d e s and performance. Journal o f Applied Psychology,1980b, 65, 139-146.
Jacobs, R ., Kafry, D ., & Zedeck, S. Expectations o f behav io ra lly anchored s c a le s . Personnel Psychology, 1980, 33. 594-640.
Johnston, J . M., & Pennypacker, H. S. S tra te g ie s and ta c t ic s of human behav io ra l re se a rc h . H ills d a le , N. J . : Erlbaum, 1980.
Kafry, D ., Jacobs, R ., & Zedeck, S. D isc r im in a b ility in multidimens io n a l performance ev a lu a tio n s . Applied Psychological Measurement,1979, 3, 187-192.
Kane, J . S . , & B ernardin, H. J . B ehavioral observation sc a le s and the ev a lu a tio n o f performance a p p ra isa l e f fe c tiv e n e s s . Personnel Psychology, 1982, 35, 635-642.
Kane, J . S ., & Lawler, E. E ., I I I . Performance a p p ra isa l e f fe c t iv e ness: I t s assessm ents and determ inan ts. Research in O rganizationalBehavior. 1979, 1, 425-478.
Kearney, J . W. The value o f b eh av io ra lly based performance a p p ra isa ls . Business Horizons, 1976, 3 , 75-83.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
Keaveny, T. J . , & McGann, A. F. A comparison o f behav io ral expec ta tion sca le s and graphic r a t in g s c a le s . Jo u rn a l of Applied Psychology, 1975, 60, 695-703.
Keeley, M. A contingency framework fo r performance ev a lu a tio n . Academy o f Management Review, 1978, 3̂ , 428-438.
Kingstrom, P. D ., & Bass, A. R. A c r i t i c a l a n a ly s is o f s tu d ie s comparing beh av io ra lly anchored ra t in g sc a le s (BARS) and o th e r ra tin g form ats. Personnel Psychology. 1981, 34, 263-289.
Krumhus, K. M., & M alo tt, R. W. The e f fe c ts of m odelling and immediate and delayed feedback in s t a f f t ra in in g . Journal of O rganizational Behavior Management, 1980, 2 , 279-293.
Landy, F. J . , & Barnes, J . L. Scaling behav io ral anchors. Applied Psychological Measurement, 1979, _3, 193-200.
Landy, F. J . , & Guion, R. M. Development of sc a le s fo r the measurement of work m otivation . O rgan izational Behavior and Human Performance, 1970, _5, 93-103.
Landy, F. J . , & Trumbo, D. A. Psychology o f Work Behavior. Homewood, 111.: Dorsey, 1976.
Latham, G. P . , Fay, C. H ., & S a a ri, L. M. The development o f behavior observation sca le s fo r ap p ra isin g th e performance o f foremen. Personnel Psychology, 1979, 32, 299-311.
Latham, G. P . , S aa ri, L. M., & Fay, C. H. BOS, BES and baloney:Raising Kane w ith B ernardin . Personnel Psychology, 1980, 33,815-821.
McCormick, E. J . , & T if f in , J . T. In d u s tr ia l Psychology. Englewood C l i f f s , N. J . : P ren tic e -H a ll, 1974.
M i l l a r d , C. W., Luthans, F . , & Ottemann, R. L. A new breakthrough fo r performance a p p ra isa ls . Business H orizons, 1976, 19, 66-73.
P e terson , N. Feedback i s n o t a new p r in c ip le of behavior. The Behavior A nalyst, 1982, _5, 101-102.
Prue, D. M., & Fairbank, J . A. Performance feedback in o rg an iza tio n a l behavior management: A review . Journal o f O rgan izational BehaviorManagement, 1981, J3, 1-16.
Schwind, H. F. An a l te rn a t iv e to beh av io ra lly anchored ra tin g sca le s : The behavior d e sc r ip tio n index. Academy of Management Proceedings, 1978, 38-42.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
52
S epler, H. J . , & Myers, S. L. The e ffe c tiv en e ss o f v e rb a l in s tru c tio n on teaching behavior m od ifica tio n s k i l l s to nonpro fessionals . Journal of Applied Behavior A nalysis, 1978, 11, 198.
Shapira, Z ., & Shirom, A. New is su e s in the use o f b eh av io ra llyanchored ra tin g s c a le s : Level o f a n a ly s is , the e f fe c ts o f in c id en tfrequency, and ex te rn a l v a lid a tio n . Jou rna l of Applied Psychology, 1980, 65, 517-523.
Shook, G. L ., Johnson, L. M., & Uhlman, W. F. The e f fe c t o f response e f f o r t red u c tio n , in s tru c t io n s , group and in d iv id u a l feedback, and reinforcem ent on s t a f f perform ance. Jou rna l of O rgan izational Behavior Management, 1978, 206-215.
Smith, P. C ., & K endall, L. M. R e tran s la tio n o f ex p ec ta tio n s: Anapproach to the c o n stru c tio n of unambiguous anchors fo r ra tin g sc a le s . Journal o f Applied Psychology, 1963, 47, 149-155.
S tap le s , W. A ., & Locander, W. B. B ehav io ra lly anchored sc a le s : Anew to o l fo r r e t a i l management ev a lu a tio n and c o n tro l. Journal o f R e ta il in g . 1972, 52, 39-48, 94-95.
Vance, R. J . , Kuhnert, K. W., & F a rr , J . L. In terv iew judgements: Using ex te rn a l c r i t e r i a to compare behav io ral and graphic sca le r a t in g s . O rgan izational Behavior and Human Performance, 1978, 22, 279-294.
Wexley, K. N ., & Yukl, G. A. O rgan izational behavior and personnel psychology. Homewood, 111: R. D. Irw in , 1977.
W illiam s, M. R. Performance ap p ra isa l in management. London: Heinemann, 1972.
Zedeck, S . , & Baker, H. T. Nursing performance as measured by behavio r a l expec ta tion sc a le s : Some th e o re t ic a l is su e s . Academy ofManagement Review. 1972, _3, 119-128.
Zedeck, S . , Kafry, D ., & Jacobs, R. Format and scoring v a r ia tio n s in behav io ral expec ta tions ev a lu a tio n s . O rgan izational Behavior and Human Performance, 1978, 7 , 457-466.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.