Upload
nguyenkhuong
View
226
Download
8
Embed Size (px)
Citation preview
UUPTEll I11
W E R I A L EFFECTIVENESS-A &SEARCH REVIEW
Campbell, Dunnette, Lawler and Yeick (1970) view
the effective manager as an optimizer in utilizing both
internal and external resources (human, material and
flnanclrl) in order to ensure the continuance of the
organlzatlonal unit for which the manager is responsible.
Emphasis in the deflnitlon is primarily on managerial
action5 and behaviour impliclt, relevant and essential for
appropriate utlllzatlon of resources. Rational evaluations
therefore require the speclficatton of the wide spectrum of
aana~erlal behavlors that are considered to be important for
g o d pcrforoatlcc. Beglnnine with the determination of the
dosaln of a manager's responsibilities, the process
continues with siaultancous identification of critical
bchavlour vltal for the appropriate use of existing and
porontlal resources. Sufflce it to say that the ai* is to
dcrtvc ncanlngful cxposltions of effective managerial
pcrlornance. Such revelations enable a more comprehensive
urdcratrndln~ of the complex inter-relationships existing
k t u e e n var lous job behaviours and organitat lonal
pcrformanccn or outcomes measured.
t h e r t a of this chapter Is to examine the
knowladge about manap?rlal effectiveness that has
accumulated over the years through research. The review of
literature is confined to large scale empirical research
falling within the ambit of global measures, predictive
studies, objective personality and interest inventories,
leadership ability tests and rater perceptions.
Global Clcasurer
Numerous studies of managerial success have
utlllrcd objective, global or administrative criteria Ie.g.
L.aurent 1961, 1962; Mahoney, Jerdey and Nash 1960).
Crltlcrl examlnatlons have indicated that global measures
such as supcrvisorv ranklngs of composite managerial
rffccrlvcnrss, raiarv or orpanfzatlonal hierarchical level
strtlsttcallv corrected for age or length of service have
srveral edvantaprs. In casrs where supervisory rankings are
in oprrrtlan onc suprrlur normally ranks a limited nuaber of
+.~brrrdlnetr aanagers. Hence the test-retest and interrater
roliahlllclc~ tcnd to be high. Further, such rankings quite
ofter, include a large sample of behaviours conditioned over
a period of tine. Conscquencly, unintentionally, @any a
tlaes the manager becomes the subject of judgement rather
thrn the urgat~l~atlonnl factors. Peer performance is used
rr standards for coaparlson of the manager's performance.
Such comparlron of the manager with his peers is appropriate
slnce all of them are usually collectively responsible for
optimizing the use of similar quantities of resources
(Campbell G . , 19701.
Overall measures of effectiveness or ratings of
success are factorially complex. Their factorial
compocitlon undergoes constant changes as a result of
complex tnteractions among the various variables (Dunnette
1963; Ucitx 1961). W i t e naturally such measures obscure
rather than reveal managerial behaviour for success. The
role of Intervening variables such as providence,
cducatlon, training etc. Is relatively unknown. Global or
adainistriltivc crlterla explain the position of the manager
,,n thc success contlnuun, but is silent on how it uas
attalned iierybcll 5 c . , 1970) .
Bchaviourr relevant to managerial success changes
ovcr timc Iliorman, 19681. It is imperative that
cffsctivcners dimensions be revleved to keep pace with
orga~~lzrclonal changes. Identical is the case of objective
merrurcr of ranagerial effectiveness. The Employee
Rclrtlonr lndex I E R I ) developed by Herrihue and Katzell (1955)
rt thc General Electric Company is a composite of eight
objective indicators of managerlal effectiveness, viz.,
absence rate, soparatton rate, dispensary visits, number of
ruggestionn subrlttcd, discipllnery suspensions. grievance
rates, w r k stoppages and extent of employee participation
in the company sponsored insurance plan. In early studies,
slgnlflcsnt relatlonships were disclosed between the
magnitude of the ERI and subjective estimates of work
effectiveness. Hwever, they were found to be insignificnnt
In subsequent studles. Consequently, the use of the ERI was
discontinued.
Dunnette and Kirchner (1958) administered the
nlller Analogies Test (WT), the Uechsler Adult Intelligence
Scale (UAlS), the Strong Vocational Interest Blank (SVIB),
the Callfornla Psychologlcal Inventory (CPI) and the Edwards
Pernon~1 Preference Schedule (EPPSI to twenty six sales
actragerr working with the 3W Co. Simultaneously the general
aanager a n d vlcc president uslng forced choice method
selected the better mrnrger out of each combination pair of
rrnaprrs. The global effectiveness score was thus the
propartlori of tlmes the aanager was selected as the more
cffrctlve of the varlous palrs. They hypothesized that the
effcctlvc manager would be more intelligent, have numerous
lntercsta and have stronger personalitles than other
mnnrlerb. Intelligence was tested by comparing the scores
of aanagers above and below the median on the XAT and those
above rnd below an IQ of 125 on the WAIS. Effectiveness
rcoras of managers 4 t h moderate to strong interest in
rtlerrt t w occuprtional groups on the Strong Vocational
Intereot Blank were compared with those of managers showing
interest on none or only one occupational group, in order to
test hypothesis relevant to interest. To test for strong
personality, the manrger's profiles on the CPI and the EPPS
were examined and divided into high and weak personalitles.
Efftctlvcncss scores vere then compared. Managers having
"fa~arrSbl&~ reports on two or all three of the variables
were compared vlth those having "unfavourable" reports on
at Ierrt tw variables. Results shoved that conbining data
[roe seprrate mcrsures yield better relationship with rated
effcctlvenrrr than if thev are considered separately.
Ca~pbell g G. 11970) observed that objective
criccrla ruffcrs frun both deficiency and contarination.
A l s o these rcruurcd only a small portion of the variance due
to indlvtdual manaperial behaviour. Variations in these
mcrrurcs w r c observed to be dependent on many, but often
krrclcvant, factors. A s these factors are not under the
dlrcct control of the mrnagcr, it is difficult to monitor
In ordcr to overcome the shortcomings of objective
crltcrir, subjective crltcrir say be resorted to (Cascio,
19781. Such rsscsneent must be based on observations of
Wrnanr at their jobs. But this practice is often
confronted with problems. Inadequate sampling of the job
behaviour domain, lrck of knowledge, lack of cooperation by
the ratets , dlfferlng expectations and perceptions of
rate- , changes in the job or job environment are some
carer of error and contamination. However, it is now
possible to overcome these by scale development methods and
training. For r more accurate picture of performance global
criteria should preferably be supplemented with systematic
observrtlonr rnd recordings of behaviour for better
undcrrtrnding of mrnagerirl success (Cascio, 1978).
The taxonorv, dcvcloped by Mcchl ( 1 9 5 4 1 , cakes
d(*tinct Ion posrlblc bctvcen the varlous types of
Instruacnt~ urcd a n d the types of predlction used in the
rrwlv of ~lanagrrial succcss. Statistical or rechanical
prrdictlonb asresb the individual on some instrument. The
acorrb obtained are subsequently correlated with a criterion
measure. Ob]rctlve personallty inventories, biographical
drtr rnd soac abill~v tests fall into this category.
Judgrmcntrl or cllntcal prediction, requires the subjective
cumblnrtlon of thcsr scores to forecast criterion status.
Qbrcrvrtlona of bchaviour at work and assessment interviews
rye cxrmplcr of cilnlcal predlctlons (Casclo 1978) .
Chiselli (1963, 1966), in his review of studies on
managerial success, reported that tests of intellectual
abll 1 ty and perceptual accuracy and personality inventories
('r's ranging between 0.25 and 0.30) are best forecasts of
managerial success. These results are attributed to the use
of criteria differing in reliability, validity and degree of
contamination. Concurrent and predictive studies were
combined and the validities were averaged. The studies
differed on many dimensions such as the relative size of
sampling errors, restricted ranges, relative reliabilities
of aeasurrs used etc. In view of these, the resultant
"crumnrrizcd" validities tended to be underestimates of the
actual dcgrrc a f rrl at ionship.
Ror-,in ( 1 9 6 8 1 rcvlcved only prediccive validity
s t u d i c r , +olclv because of less arbiguicy in the
intcrprrtntior, c r f rc.sults. Verbal ability tests, used to
srnsnrc. inrc.1 1 tgcnce, were found to be good predictors of
f Lrst Itnr su;>crvlstiry pcrfar~anct, but not perforcance at
the high'-r Icvcls. because of restriction imposed by the
range probl c q m .
Crlmsrlv and Jarrrtt 11973) used a matched group,
conerjrrc?~~;-val id1 tu design to detercine the extent to vhich
rerrtal rbi l i c y test scores and scl f description inventory
rcorar obtained during assessment prior to employment
dlstingulsh between top and middle level managers. The
matched group design was used to control two moderator
variables (age and education) which were posited to be
related to test performances and managerial achievements.
Our of 037 appllcants 100 subjects were selected. Each of
the 50 top managers was paired with one of the 50 middle
manap,crs considering age and college education. Success
crltcrlon was based on the level of managerial
rcrponsibillty attained prior to assessment. An advantage
of the design Is that the data was gathered under actual
raployernt eondlt ions. On the contrary, the disadvantages
arc the large wa5rage in subjects used and the difficulty in
lnccrprctcrtlon of results because "retching the variables
contrallcd inevitably makes one of the matched sarples non
rei~rcrr-nt a t l v e of rhr population from which it was drawn"
((a*cl,,. 3978, p . 2 3 7 1 . lnspite of this probler, Crimsely
and Jarrvlt chuse to coapr~mise representativeness for
inrrra~c*d ccr~ltrol of these two moderators which affect the
rclatlonship bctwrcn t e s t scores and managerial achievement.
AT d T 1 3 management progress study was begun in
1956 to investipatr the career development of persons
enploycd by tire orpanlrrtlon. A nueber of significant
reports of rerul t a (Rcrlev, 1965; Bcrlew 6 Hall, 1964, 1966;
Bray, 1962, 196(4, 1966; Bray and Grant Berlev, Rychlak and
Katkovrky, 1965) h a s been made.
Bray ( 1 9 6 2 , 1 9 6 4 ) undertook t h e s t u d y t o t r a c e t h e
c a r e c r b of aanage r s from t h e t ime of t h e i r e n t r y i n t o Be l l
System. The s t u d y in t ended t o o b t a i n an a p p r o p r i a t e
psycholo' lcal d s w s s n e n t of a number of ncv young e a p l o y e e s .
I t a l s o a t t empted t o s t u d y them p e r i o d i c a l l y th rough ou t
r h e l r c r r c c r . Paper and penc i l t e s t s , work samples ,
projective C C ~ ~ S , c l i n i c a l i n t e r v i e w s and p a r t i c i p a t i o n i n
group p r o b l c s s and l e a d e r ' s group d i s c u s s i o n s were used a s
a s s c s s w n t a t - a su rcs . 5 t u d l e s t o a s s e s s s u b j e c t ' s l i f e , work
e n v l r o n u e n t , t r a i n l n p , s a l a r y , t r a d i t i o n s and e o a l s of h i s
o r g a n l 7 r t l o n , and t h e a o r a l c of t h e employees' work group
v e r c ln t~ .nd~ . i i . F o l l o b up of sen who had t e r m i n a t e d t h e i r
c s p l o v e u r ~ t i l t h A I A 1 t o p e t h c r with cep loyees c o n t i n u i n g a t
tht* Rcll \ v \ c c r w a s a l s o under taken.
I'ilirr hundrrd dnd twenty two r e n employed i n 6
cimpanl I + v r r r a+cths+etf over scvc ra l y e a r s . Most of t h e s e
r t u d l c * rt-1 t t cd t o t h c n a t u r e of judgements nade d u r i n g
~ s s t . s % a t % r ~ t , d11tl i t 5 r r l a t i o n s h i p t o l a t e r job per fo rmances .
Ihr i n i t i a l q t c y i n such a s s c s s a c n t p rocedures c o n s i s t e d of
Idrrrt l f v l t i p c h a r ~ c t r r l s t l c s t o bc a s > c s s c d . Out of t h e
nuarrouu l t r a s i t i r . ~ ~ t i f i c d , a l i s t of twenty f i v e qualities
w a s s r l ~ c t r d . f h c c r v a r i a b l e s vc re s t u d i e d by t e c h n i q u e s
I t k o I n - b j s k r t n c t h t d , i n t e r v i e u , manufac tu r ing problem,
g roup d i s c u s . s i ~ ~ n , jrrcljcctive t r s t r , p r p r r and p e n c i l t e s t s ,
and questionnaire~.lnfor.rtlon was also obtained through a
personal history questionnaire,an autobiographical essay and
seventy Q-sort self descriptive items.
Interviews were conducted by individual staff
memkrr. Tests, questlonnaIres and other exercises vcre
rdmlnlstered in groups. The group consisted of six
participants at a time. T w staff eembers recorded their
asseas~ent of each participant and evaluated them
separately. The in-basket technique was revieved by one
staff membor. One observer for the group exercises
evaluated performance of the individuals in these exercises.
Ratlngs of peers and observers was also considered. This
vas revlewd and an lndlvidurl report was prepared by a
trained prychologlst. The paper and pencil tests w r e also
scored. For Q-sort itees the employee identified those that
uere "most" and "least" like him. The five point scale was
used to rate each indlvldual on twenty five characteristics
vit., organization and planning, decision making,
creativity, human relations skills, behaviour flexibility,
p.rson@l Impact, tolerance of uncertainty, resistance to
stress, scholastic aptitude, range of interests, inner work
standards, primacy of w r k , oral communication skills,
p.tcrptton of swlrl cues, self-objectivity, energy, realism
of e~pccations, Bell System value orientation, social
objectivity, need f o r advancement, a b i l i t y t o d e l a y
g r a t i f i c a t i o n , wed f o r s u p e r i o r approval , need f o r peer
approval , goal t l e x i b l l i t y and need f o r s e c u r i t y .
Af te r completing t h e r a t l n g s on a11 v a r i a b l e s t h e
i n d i v i d u a l ' s p o t e n t i a l a s a management person i n t h e 6ell
System was evaluated by t h e s t a f f . P ro jec t ions of an
employee's continuance a t t h e Be11 System and the l i k e l i h o o d
t h a t he would achieve middle management p o s i t i o n and
conc lus ions of whether he should be promoted thus was made.
These - a t l n g s were discussed and r e v i e w d by s t a f f members
vhereever i t was considered necessary.
Resu l t s of these s t u d l e s provlded evidence i n f o u r
a r e a s v l z . , f a c t o r s underlying t h e 25 assessment v a r i a b l e s ,
t h e r e l r t l v e :ontr ibut lon of t h e s e assessment methods a d
t e a t s t a r a t l - - 8 s on d i f f e r e n t assessment f a c t o r s and t h e i r
o v e r a l l prcdic ion c r p a b l l i t y , r e l a t i o n s h i p s b e t w e n t h e
s t a f f ' s o v e r a l l p red ic t ions and l a t e r promotions and s a l a r l y
l e v e l s , r e l a t i o n s h i p s between e a r l y job cha l lenges and
performance and l a t e r c a r e e r success .
The r a t i n g variables used i n t h e f l r s t s tudy
d l f f e r o d from those i n subsequent assessments. For
purposes of f a c t o r a n a l y s i s , t h e employees were d iv ided i n t o
those wlth c o l l e a e degrees r t t h e t ime of e n t r y i n t o s e r v i c e
(N.207) and t h o r e without degrees (N-1081. Cor re la t ion
matr ices were d r a m up f o r each of t h e groups. Uherry's
(1959) h i e r a r c h i c a l method was used f o r f a c t o r ana lys i s .
S imi la r f a c t o r s vere obtalned f o r both samples. General
e f f e c t i v e n e s s , admin is t ra t ive s k i l l s , i n t e r p e r s o n a l s k i l l s ,
con t ro l of f e e l i n g , i n t e l l e c t u a l a b i l i t y , vork o r i e n t e d
motivation, p a s s i v i t y and dependency were t h e e i g h t common
f a c t o r s .
Correlations between various t e s t s and &ssessment
methods and s c o r e s on each of the v a r i a b l e s i d e n t i f i e d by
t h e f a c t o r a n a l y s l s revealed t h a t some methods vere more
helpful t o the s t a f f f o r making judgements of indfv idua ls
concerned.
Bray and Grant (1966) determined t h a t t h e
behrvlour on t h e In-basket , performance i n manufacturing
problem and group exercises accounted f o r f i f t y percent of
t h e var iance In t h e overa l l s t a f f p r e d i c t i o n s f o r t h e
c o l l e g e sample and t h i r t y per cent of t h e var iance i n t h e
non c o l l e g e sample. M n t a l a b i l i t y measures accounted f o r 6
per cent and 1 2 per cent of v r r lance i n t h e two samples. In
Ju ly 1965, f i v e companies provided t h e progress made by t h e
rubfec ta i n t h e s tudy. I h e management and s a l a r y l e v e l s
when compared with p red ic t ions made a t t h e time of
assessment r h o w d chat 82 per cent and 75 per cen t of t h e
college and non-college educated subjects had been correctly
identified as likely to achieve middle management level. 94
per cent were correctly identified as those vho vould not
progress beyond first level of manrgement. Situational
techniques were found to predict success to a considerable
extent. Cognitive and personality measures had limited
scope.
The Hanagement Progress study data was used to
study the effects of early job challenge on later managerial
effectivenets (Berlew, 1965, Berlew and Hall, 1964, 1966).
Forty f w r college men continuing on their jobs with the
company were studied. An estimate of the individuals degree
of job challenge during his first year with the company vas
prepared based on Information elicited through interviews
with company officials.. From this information, Berlew and
Hall estimated company expectations for each Individual in
elghteen bchaviour areas. They are technical competence,
loarnlng capacity, imagination, persuasiveness, group
membership skllls, communication skills, supervisory skills,
declsion making, organizing ability, time-energy commitment,
racrlflce of autonomy, soctability, acceptance of coapany
norms, self development, maintenance of public image,
loyalty, productivity and initlrtive.
Expectation for each indlvldual was rated from one
(low) to three (high). Job challenge was measured as the
total of all ratings held by the company for each person.
If expectations e r e high it was assumed that challenge was
hlgh. Data were collected to determiw relationship between
ruccrss and job challenge. In 1962 a global appraLsal of
each man's overall performance and potential was combined
ulth an index based on present and starting salaries to
constitute a success index. A number of measures of job
performance was also lncluded to form a performance index.
Analyses revealed that persons judged to have more
challenging first fobs were found to accomplish more than
chore ulth less challenging first jobs after four or five
years in thc company. The foregoing results indicate that
first ycsr job challenge is reasonably associated with later
judgements of success and job performance.
The group cxcrcises end the performance on the in-
barket tesc were found to be the most llrportant determinants
of the staff member's assessment on most variables. The
predict lvr val ldi ties of the global predictions w r e
moderately high. With the establlshment of the personnel
assessmant program on an experimental basis an increase in
proportton of persons with better performance and high
potential into first lIne management positions was observed.
The promotional patterns of one hundred and
eighteen college men hired during 1956 and 1957 and
contlnuing to work at the Prudential Insurance Company was
studied by Selover (1962). At the time of hiring,
information about the schools attended, course of study,
activities undertaken at college, w r k and military
experience, college grades, scores on tests of verbal and
arithmetic reasoning ability were obtained. Of these,
scorer on arithmetic reasoning and vocabulary tests,
rcholastlc achievement, part-time work and extra-curricular
partlciprtion, leadership abilities and number of jobs held
before joining Prudential insurance Company were found to
have a strong impact on advancement. There five measures
w r c combtned into a single advancement potential score.
The general applicability of the scores for
prcdlctLng promation was experimented on four additional
groups. The first group consisted of thirty employees hired
betreen 19L8 and 1951, the second eighty five employees
hired between 1952 KO 1955, the third thirty six employees
hired betwen 1958 and 1959. These employees were
continuing with the company in 1961. In addition, the
advancanent potential score was computed for 48 men vho had
been r s k d to resign and 25 men who had been promoted to
hL8h level jobs. The phi co-efficientr of 0.36, O.tO, 0.35
and 0.65 indicated that the scoring system was moderately
rr1at.d to promotions in all four cases.
The rcrultr support the view that aeholastic and
non-scholartic factors facilitate prediction of career
succera for managers. Selover (1962) opines that college
graduates who were promoted rapidly hid better on general
ability tests, achieved high academic grader, war active in
coltcge, displayed greater leadership abilities in college
or military service and joined the company soon after
gradu at ion.
Husband ( 1 9 5 7 ) reported results of his study of
career success of 368 graduates of Dartmouth College. Large
aedlan dlfft*rences In current income between scholastic and
nonscholast i . men were observed. Men who were academical1.y
g o d , prrtlclpated in extracurricular activities, and
demonstrated leadership potential were found to be more
successful in their careers as measured by promotion and
salary Ievclr.
One hundred and ninety six former MBA students of
the Stanford Univcrslty Graduate School of Business were the
subjects of the study conducted by Williams and Harrel
( 1 9 6 4 ) . They found that the current salary after adjusting
for leneth of years or:t of school was moderately related to
grader received at the school ('r's of 0.13, 0.14 and 0.22).
It w r a also related to overall school accompllshnent
(r-0.18) and to lcadcrship in campus (r-0.24).
Harrell (1967) observed that the most suitable
criteria of job success were Hemphill's Position Concern
Quertlons, present compensation, present job success as
perceived by the MBA graduates and the number of hours in
the w r k week. Slnce inventories such as the Public Opinion
Quertionnairc, Lnitiating structure and decision making on
Ghlselll's Self-Description Inventory, Undergraduate Grade
Polnt Averate, Individual Background Survey, Ascendance on
the C - Z Temperament Survey, initiative on the Ghiselli SeLf-
Description Inventory and nanlc Scale on the HnPI yielded
posttivc results, its use was recomaended for admission to
Graduarc Sch>ols of Business.
Ncycr '1963, 1965a, 196Sb3 studied the utility of
tert mcrsures r,hcn other factors related to promotion are
held relrrlvely constant. Hls findings are based on a
~erie9 of studits conducted at the Jewel Tea Company in
1959. Ihc purpose of the study was to identify managers for
pronution. Lach marager was assigned a promotability rating
on a six-poln~ ratlnr acale. A follow up study was aade in
1965 of one hundred and seventy eight managers assessed
during 1959 and 1960. Assessment ratings, biographical
factors and test scores were related to promotions during
the Lntervrnlng years. The assessments shoved a strong
relrtlon~hlp to later promotlonn. Of the eighty five
persons identified for promotion, sixty seven vere actually
promoted. Of the ninety three persons not considered
suitable for promotion by the assessment only twelve vere
promoted. Since Meyer anticipated chat the assessment
ratings vould have had an impact on the actual promotion
declrlon, he Lnvestigated the role of other factors. Such
factors as age, education , and initial management level
were found to be soderately related to promotional
rxpcrlcncc. Each of these factors and the assessment rating
vcre then wclghted equally to yield a "promotability
probabll icy index", vhich vas found to be strongly related
c u proaotton.
The study nighlighted the fact that promotion
polLcLcr in the Jewel lea Co. from 1958 to 1965 gave
Laportrncc t o ranagcrs uho vcre relatively younger, more
educrtcd. had begun their career at lower levels and w r e
vorkln8 in tht* buvlnf, serchandising or adsinistrative
deparLucrits of the coapany.
Thc dcrrcc of rclationshlp between the various
cognlclvc and pcrsonallty tests used in the assessment
proF;r&ta a d proeotlon uus detcrained. The other influential
factors were held relatLvcly con-tant. The test information
for promoted and non-promoted persons with the same or
rlmtlrr ,corer on the "prunotabillty probability index" was
perused. The promoted managers tended to have higher scores
on the general intelligence test than the non-promoted
eanagerr. lkyers observed that those vho were promoted were
better delegators, more flexible, less manipulative of
people, not constantly dependent on superior's approval and
were more forthrlght and emotionally well disposed.
Test information was related to various
effectiveness measures by Albright (1966) . One hundred and
clghty three field sales managers working with the American
Oil Company were taken as sample for the study. Predictive
relationrhips were based on tests administered to nearly one
thlrd of the men i r ~ 1955, who were then working as salesmen
In the fir.. Conc~rrent relationships, on the other hand,
were bared on tests administered to eanagers shortly before
cffectivr~~css rattngs ware obtained. Global performance and
prolrotabtllcv ratings made by superiors, and salary growth
r a t e were the acasures for effectiveness. Correlations
between a l l test scores taken together and each of the
crltrrla w r c comptlted. Coreputations for the total group
anJ for subgroups designed according to age above and belov
forty, rnd according to functional areas were made.
in rome Instances low or moderate concurrent
rclrtlonrhlps (ranglng between 0.15 and 0.101 were obtained
&twr\an glabrl performance ratlng for vocabulary knovledge,
measures of dominance, self sufficiency, decisiveness and
SVIB scales. Moderate predictive relationships (ranging
between 0.25 and 0.15) was observed for global performance
rating for SV18 scales concerned with positive interest in
raler, advertising, verbal and business occupation and
negative interest Ln skilled trades and outdoor occupations.
Global promotability ratings for intelligence and vocabulary
seasurer, pol it lcal and business oriented values,
personality measures and SVIB pertaining to business verbal
and rdeinlstrative jobs shoved low to moderate concurrent
and prcdicrfve relacionships (0.15 to 0.35). Concurrent
rclatlon5htps ucre greater for men over forty than for those
undcr f u r t y .
Studics vcrr carried out at the North American
Avlstlofr [Trnopyr, 1960, 19618, 1961b, 1961~). to assess the
valldlt$ of test+ such 8s chc Otis Mcntal Ability Test,
supurvlhurv knkrblrdpe tests, the fhurscone Test of Mental
Alertness ctc. for prcdlcting success as measured bv salary
advrncrmrr~t , pronot ion, perforrance rating etc.
Currelatlanr obtained were uniformly near zero. The studies
shoved c h a t none of the tests was capable of assessing
supcrvlsary cffrctivencss.
Later studlcs conducted at the same organization
shoved w r c paatrivc results (Tenopyr 1962, 1963, 1965;
Tenopyr and Ruch, 19651. Tests showed that successful
managers w r e brighter (r-0.66 with Otis IQ) and more
proficient in mathematics. Verbal comprehension and
Leadership Evaluation and Development scale yielded moderate
correlation8 of 0.29 and 0.36 respectively. Possibility of
curvilinear relationshlps were checked out by computing etas
between test8 and rrlary criterion. Belationship with
verbal comprehension showed no significant change, but the
relationship between salary and human relations was
curvtlinear ( r- 0.36; eta - 0.16). Utility of examining
curvilinear relationships between predictors and criterion
mersures war thus establlshed.
Albrecht, Claser and Marks 11960) studied thirty
one district nrrketing managers who soon after promotion
mere given r eentence completion test and a huaan relations
test, i~telllgence and crittcal thinking tests besides
partlcipiting in an interview. They w r e also asked to
complete a qucstionnrlre on their personal history. Three
psychologlstr ranked the subjects according to predicted
rffectiveners in four areas viz., forecasting and budgeting
effectiveness, rrles performance, interpersonal relrtion-
rhipr and overall ptrforeance. After completion of a year
on the job, the8e manager8 were ranked on the above factors
by tw #uperlors and peers. There three sets of rankings
w r e then combined to form a co~porite index for each
tndividurl in each area. The psychologists' predictions
were then correlrted vith the composite index scores.
Scores on the tests were also correlated with the composite
rankings.
The comblned information was found to yield better
predtcrlons of later effectiveness of these subjects than
any test taken individually. This difference was least in
case of forecasting and budgeting effectiveness and most for
prtdlctlng sales performance.
Flanrgan and Krug ( 1 9 6 4 ) administered the SCORES
battery con~isttng of 38 variables distinguished into 6
broad areas viz., supervlslon (51, creativity ( C ) ,
or~anlzatlon (01, research ( R ) , engineering ( E l and
srltrmanshlp ( S ) to sixty engineers wrking with Lockheed
Aircraft Corporation. They were divided into four
homolgeneous subgroups as equally matched as possible on age,
sducatlon and company experience. They were differentiated
into supervisory and non supervisory categories and rapid or
slow promotabll icy. Hlgh and low promotion categories
scored ldrntlcally on creativity and engineering. High
promotion manrgers scored higher on supervision,
orgrnitatlon, research and salermrnship. Engineers in
rrnrgement jobs dtffered from those in technical jobs.
Flanagan and Krug were able to assign participants into the
four groups based on the individual's SCORES record. The
study eupports the view that a battery of tests aimed at
deteralnlng functlonal area aptitudes and traits will be
capable of identifying "general" effectiveness in management
and for making placement decisions.
Objective Personality and Interest Inventories
A number of investigators (Cuion and Cottier,
1966; Hedlund, 1965; Korman, 1968; Nash, 1965) have
obtained ncgatlve results on using personality and interest
seasurer In forccasting managerial effectiveness. However
Chlsclll's rcvlcw (1966) has been positive. The aatched-
group d e s l ~ n used by Crlmsely and Jarrett (1973, 1975),
yielded much higher coefficlents from several standardized
personrllty oeasurcs used. A questionnaire comprising the
gcncrol actlvitv, sociability and emotlonal stability scales
based on t h e Cullford-Zlomerman Tempera~ent Survey yielded a
multiplc R of 0.52. The General Activity scale when
combined with the verbal reasoning and numerical ability
scales of the Lwployec Aptitude Survey yielded a multiple R
of 0.61. These coefflclents denoted concurrent validity.
R e ~ u l t s verc ascribed to the severe restrictions imposed by
the m a t c h 4 group experimental design and the fact that the
data were gathered under employment rather than research
conditions.
The Early Identification of Management Potential
(EIMP) study, is a ploneer in the validation of tests and
biographical inventories against measures of managerial
tffectlveness. The management of the Standard Oil Company
of New Jersey (SONJ) supported this research in order to
shed llght on the measurement of success and identification
of persons with potential for success. Results of the study
arc reported by L.aurent (1962).
F Iur hundred and forty three managers working for
SONJ and five of Its affiliate companies vere the subjects
of the study. These subjects vere drawn from all the
different functlonal areas. Three measures of managerial
c f f ~ c r f v r ~ i ~ ~ s were chosen - position level, salary history
roll effcctlvrness ranking. These measures were then
~uat~lncd to form an overall success index. Factor analytic
methods ensured i t s independence from both age and length of
oxparlrncc In the firm. This measure of relative managerial
effectiveness, did not spccltically identify or quantify
directly, actual nonagerial behaviour. Hovever, it
reasontbly differantlatcd between managers in their
utllisaclon of organltatlonal resources in their careers.
Each aancrger completed a questionnaire about famiIy,
education, vocational planning, finance, hobbies and leisure
tlme actlvitles, health history and social aspects. Also
included were measures of verbal ability, inductive
reasoning, oanagement judgement, managerial attitudes and
prrsonallty measures in the Cuilford-Zlmmerman Temperament
Survey. These test scores were correlated with the overall
success index. The degree of relationship between the
overall success index and items in the biographical survey,
oanagement judgement and Cullford-Zimmerman inventories was
examined. Double cross validation was resorted to assure
scablllry of relationships discovered. The total sample was
dlvlded into two groups of 222 and 221 managers each.
Scorlng keys developed on each group was cross validated on
tho other group. Next the test and responses yielding
highest and most uniform correlations with the overall
success Lndcx w a s combined to yield a single composite test
acorr. The correlation between the overall success and
comporitr rcorcs was 0.70. Four measures emerged as useful
for the evaluation of success. They were the Cutlford-
ZLmmeroan Survey, the Hclnagement Judgement Test and the two
measures of intelligence. Evidence from the data was
suppartivr of the quallttes identified in the study as those
related to 8cneral potential for management, regardless of
functional, company or job differences within the
organltatlon. It war further contended that the spcial keys
and weighting methods developed Ln the study could be
applied across different companies and even different
countries with some but insignificant loss of
predtctivcners. To Chat extent the rtlsdy is regarded as
ruccerrful in identifying personal or trait correlates of
managerial effectiveness.
Another significant large scale study of the
personal correlaccs of managerial effectiveness was done by
the Industrial Relations Centre of the University of
Minnesota (Hahoney, Jerdce and Carroll, 1963; Mahoney,
Jerdec and Kash, 1960; Hahoney, Sorenson, Jerdee and Nash,
1963l.Four hundred and fifty two managers fro. thirteen
divcrrc ftrrs in Hlnnesota constituted the sample for the
study. tach v l the participants was asked to complete a
brtrery of t e s t s consisting of the Uonderlic Personnel Test,
€.apathy 1 c ~ s t , St rong Vocat ional Interest Blank (SVIB) , Lalifornia Psvchological Inventory lCPll and a Biographical
questlnnnalrc. In total ninety eight predictors were tried.
Managerial rffectlvcness was measured by a panel of top
company officials who ranked the participants on the basis
of management competence. The results showed that the
ranking scorcs were independent of manager's age but related
to organizrtional level In some of the companies. The
ramplr war spl i t randomly into two. Scores of the various
t e s t s *re r tud ied on t h e b a s i s of t h e i r capac i ty t o
d l s t l n g u i r h between t h e seventy f i v e top and bottom ranked
managerr i n one of t h e groups. The measures t h a t w r e
r t a t i a t i c a l l y r l g n i f l c a n t were used t o develop a s c o r i n g
ryetern which war c r o s s v a l i d a t e d on the managers i n t h e
o t h e r rample. Resu l t s obtained i n t h e c r o s s - v a l i d a t i o n
groups from applylng p a t t e r n s of p r e d i c t o r s and c u t t i n g
s c o r e r rhoved t h a t 29 per c e n t and 62 per cent w r e judged
l e s s and more e f f e c t i v e , and received scores above t h e
c u t t i n g score on p r e d i c t o r combination. 71 per cent and 38
per cent scored below t h e c u t t i n g score and w r e judged l e s s
and r o r c e f f e c t i v e r c s p c c t i v e l y . The Uonderllc Personnel
Test and domlnancc measured by C a l i f o r n i a Psychological
inventory uere s l g n l f l c a n t . The Strong Vocational I n t e r e s t
Blrnk showcd t h a t e f f e c t i v e managers tend t o have s i m i l a r
l n c e r c s t s as o ther i n bus lness . Their i n t e r e s t s were found
t o d i f f e r from those i n a g r i c u l t u r e and s k i l l e d t r a d e s . The
Biogrrphlcal quee t lonna l re showed t h a t ef f e c t i v e managers
were more educated, had b e t t e r educated spouses who res igned
t h e i r jobs s h o r t l y a f t e r marr iage, p a r t i c l p a t d In s p o r t s ,
hobbles and were r e l r t l v e l y young a s corpared t o t h e l e s s
e f f c c t l v e managers. E f f e c t i v e managers uere a l s o found t o
br more l n t e l l l g e n t and dominant than t h e l e s s e f f e c t i v e
onar .
Kirkpatric (1961) has described some of the
results of the study conducted to determine the biographical
correlates of job effectiveness for the executives of the
American Chamber of Commerce. He developed a biographical
questlonnairt based on interviews with eighteen Chamber of
Commerce executives. It consisted of a11 background
Lnformrtion such as early history, family background,
education, employment history, military service, physical
conditionr and social inforartion. Six hundred executives
of the Chamber of Commerce responded to the questionnaire.
f'lanagrn's ( 1 9 5 1 ) listing of the critical requirements of
Alr Forcr Officer's jobs was used to form a list of
effec~lve exccutlvc behaviours. This was used to group each
of the respondents as above average, average or below
average i n eeetinp the critical requirements identified by
tlrnagan. Adapting the "forced" choice technique' fifty
perccqnt trf the subjccts were grouped "average" and twenty
live pcrccnt each were terred "above average" and "below
averrge". Tuo groups conslsttng of fifty members each
dcr1gna;c.d as above and below average, and who were matched
on age and length of job experience and rated high on
effscclvrners were chosen. Comparlson between the responses
of these tw lroups ylelded thirty five items. These were
c0ebin.d a d scored to enable prediction of above average
ptformancr. The rcortng system was agaln cross-validated
on two groups oE thirty below average and thirty nine above
avcragc executives. The point biserial correlation obtained
was 0.56.
Kirkpatric 11961) concluded thac the successful
Chamber of Commerce executive has a middle class socio
tconoaic background. was well educated, engaged in many
extra curricular actlvlties, emphasised communication skills
and entered the Chamber of Commerce relatively early in
life. However some of the biographical variables tend to
dcscrlbc s u c c e s ~ rather than predict it.
In the Sears Roebuck investigations an executive
brttrrv of standardized tests was used to predict criteria
of *anoyrrial success such as promotion rates, nomination of
man.rgcr4 blth high succcss potential and high employee
acritudr %:*ale results. Bcntz (1963, 1967) has reported
rc4ul tz ;>t,rlalnlf\g to the study. The Sears organizations
uartl four ~ c b t s namely Amcrlcan Council on Education Test,
tull!ord-!4arcl1~ Invcnlorles, Allport-Vernon Survey of Values
4rrci Rudcr f'reiercnce Record to predict executive
cffrctlvcnr\s. The tests were correlated against employee
morale sctrrcs. Analyses e r e aadc using discriminant
functlon and multlple rcgresslon.
Multtplc correlatlans approaching 0.75
f # I l L n ~ below O.LO *re reported. The results 1
objective personality and interest invun+orler can predict
managerlal success to a considerable extent. Moreover ,"a
cluster of psychological characteristics contributes to
general executive competence that transcends the boundaries
of rpecialized or non-speclalired assignments"(Bentz,p.118).
Horse and Wagner ( 1978 ) developed an instrument to
measure and evaluate managerial performance. The process of
managerlal performance was the focus of the study. A paper
and pencil instrument was developed to determine
rimilarltles and differences among sanagerial jobs.
In the first stage of the study, specific
behavlour and activlcies characteristic of managerial work
was ldenrlfled. Based on literature reviews and interviews
with six top corporate executives in six different
comprnlcs, one hundred and six item statements vere
devclopcd. They were then clustered around nlne roles vit.,
st rrte~lc probl ea solving, resource hand1 ing, conflict
handling, organlrlng, information handling, motivating
provldfng for ~ r o v t h and development, coordinating and
managing chc organization's environment. This instrument
was adstnirtered to one hundred and flfteen managers. The
Lnstrument was aodlfled based on suggestions made by thtm.
this enabled the reduction of items to ninety six, clustered
around the nlne roles. Four hundred and six managers
completed the modified questionnaire on another manager.
The Crovne and narlowe (1960) Social Desirability Scale was
also completed by the participating managers on themselves.
This modified instrument was factor analyzed and a
principal component solution was extracted. Six factors
with eigen values exceeding 1.00 were obtained. Based on
rcnults obtained the instrument was reduced to contain fifty
one items. Correlations between item score and total score
exceeded 0.40 and correlation of overall score with Croune
and narlovc scale was 0.11 only.
The final fifty one item instrument was factor
rnalyxcd uslne, correlation matrix obtained from the 406
managers. Six factors emerged with eigen values exceeding
1.00. The frctors in descending order are - managing the
orgrntzation'i environment and its resources, organizing and
coordinating, informrtion handling, providing for growth a d
development, motlvaclng and conflict handling, and'strategic
problem solvlng. Orthogonal rotatlons of the six factors to
the varlmaw crlterion resulted in 56 per cent of the
variance being ilccounted for. Analytic obliqw rotations of
the nix factors was resorted to to improve the structure and
test for lnterrclatedness. Moderate relationships were
obtr1n.d. Rerults suggest that insplte of the distinctive
character of roles, they interrelate to describe
slailaritieo and difference# in managers' jobs.
Both content and concur ren t v a l i d i t y of t h e s c a l e
war e r t a b l i r h e d by t h e d a t a . I n t e r n a l r e l i a b i l i t y war found
t o be r a t i r f a c t o r y . Repl ica t ion of t h e r tudy y i e l d e d
s i m i l a r r e r u l t r . The 51-item instrument war t r i e d on a new
sample of 620 managers. These managers represented a number
of f u n c t i o n s , l e v e l s and experience and were d r a m from two
o r g a n i t a t l o n s . The replication again i d e n t i f i e d s i x f a c t o r s
with eigen values exceeding 1.00. Oblique r o t a t i o n s
re ru1 t .d i n a f a c t o r p a t t e r n matr ix s i m i l a r t o t h e o r i g i n a l
r tudy . Factors were found t o be moderately i n t e r r e l a t e d
( h i g h of 0.64 t o a low of 0.29).
The r c p l l c a t f o n d i f f e r e d from t h e o r i g i n a l i n two
a s p e c t s . F i r s t , the variance accounted f o r was 53 per c e n t
1 . . 3 per cent l e s s than t h e f i r s t c a s e ) . Uhile f a c t o r
one accounted f o r 17 per c e n t , f a c t o r s i x accountrd f o r o n l y
5 per cent of the variance. However i n t h e r e p l i c a t i o n
r t u d y , f a c t o r th ree accounted f o r 12 per cent and f a c t o r one
only 6 per cent of t h e var iance accounted f o r . Second, item
s ta tements loading on a f a c t o r remained uniform but t h e
a c t u a l f a c t o r loadings d i f f e r e d . The two a s p e c t s d e n o t e
t h a t t h e varlance assoc ia thd with each f a c t o r cannot be
& r n e r a l i z e d and wil l vary according t o t h e s i t u a t i o n .
t r p o r t a n c ~ of the item i n d e f i n i n g a f a c t o r a l s o cannot be
tOnera l lg& and depend8 on t h e r p e c i f i c s i t u a t i o n ,
~lultlple regression was used on the data from 2
different organizations to teat the influence of the
situatlon on effectiveness. In the first case the
significant factors associated with managerial effectiveness
were information hand1 ing, and strategic problem solving,
while in the second it was managing the organization'e
environment and its resources, organizing and coordinating,
and motlvatlng and conflict handling. The discrepancy in
results lndlcates that the importance of each item statement
in defining the factor loadings varies according to the
speciflc managcrlal situation. Based on personality,
predlrposltlons, managerlal style, skills and abilities of
managers some particular roles are emphasized while yet
others are ignored.
Reviews concurred on the point that there is r
great MA to focus attention on the various situational and
indlvldual difference variables which are likely to affect
the prdictrblllty of managerial effectiveness. Job
khavtoura which contribute to managerlal success or failure
differ across hierarchical levels and organizations.
According to evidence obtained by Cuion and Cottier (19661,
custom mrdr scoring keys validated for rptclfic
or&anlratlonal settings are more useful than standardired
prrronrl i c y orasurcs with st andard scoring systeas. Scaler
developed to predict occupational choice (e.g. the Strong or
the Kuder) or Prychological Alienation Tests (reg. the WPI)
are seldom capable of predicting managerial behaviourr such
rr effective planning or delegation (Campbell fi., 1970).
h d o r r h i p Abil lty Tests
kasures of leadership ability have proved to be
p o d predlctors of managerial success since they are
direct1 y relevint to managerial job requirements. Scales
derlgned to measure two major constructs underlying
managerial behrvlour - conslderation and initiating
rtructure have been developed and used in many situations
(Flelrhean, 1 9 7 3 ) . In one revleu of 7 studies which used
the Lcrdcr Opinion Questionnaire (LOQI, Korman (1966)
reported a median value of -0.02 for validity coefficients.
Howvtr, only 2 of these studies used industrial samples a d
w r c predicrlvc tn nature. Another revlew by Korman (1966)
of studles uslng the Leader Behaviour Description
~ s t i o n n r i r e ( L B D Q ) , reported higher validities. All these
studies w r e concurrent, and hence it was imporsible to
detrrmlnc whether lnltlatlng structure and consideration
n r e cruses or ef fectr of managerial effectiveners.
brcrlption of managerial behavlour by relevant others are
.ore accurate aa predictors of managerla1 success than are
self dercriptions. Kerr c. (1974) have observed that the use of LOQ, LBDQ and SBDQ (Subordinate Behaviour Descriptton
Qu6stionnaire) yields significant concurrent relationships
between leader behaviour predictors and various criteria.
Such enhanced predictability, could be attributed to
researchers' efforts to conceptualize and measure
situational variables relating to leadership behaviour and
infrequent use of averaged data. In such cases analysis is
dona by the vork group rather than by the individual.
Frequent use of moderator variable research strategies was
observed. A requisite of this is the independent analysis
of the lndivldual. On the other hand the causal
implications of consideration and initiating structure are
unknown. The prcdictivc validity of the scales relains to
be proved a s yet (Kcrr and Schreinsheia, 1 9 7 4 ) .
The relation between ratings of leadership
effectiveness and conflict resolution methods was
investigated in a large midwestern manufacturing firm
(Oernardln and Alvares, 1 9 7 b l . Hypotheses based upon the
managerial grid developed by Blake and Mouton (1964) were
tasted. The team, the middle-of-the-road and the task
oriented supervisor did not prefer the confrontation,
compromlra and forcing strategies respectively. No relation
k c w e n mrnagerial orientation and leadership effectiveness
ratings of first line supervisors was found. Supervisors
did not seem to differ in their preferences for conflict
res0luti0n strategy as a function of managerial orientation.
Perceptions of resolution strategies vere observed to differ
as a function of organizational level and conflict type.
Thus the rerults indicated that the grid placement was a
poor predictor of either perceived managerial effectiveness
or conflict resolution methods employed. In a rejoinder to
these obrervatlons Blake and Mouton (1976) had refuted the
use of self assessments obtained prior to a seminar to test
hypotheres about managerial effectiveness or conflict
resolving methods, as the data are unreliable over estimated
judgesents by the participants.
Rater Perceptions
Since research has repeatedly proved the necessity
to consider ratings of different constituencies, the use of
appralral techrrlques by each of these groups of raters is
exasined hereln. A prime criterion for evaluation of
success is that appraisal scales should be used only by
p.oplr who are conversant vlth the aims and objectives of
the job; people who usually observe the ratee performing on
the job; and pople who are capable of making reliable and
valld rrtimrtrr (Latham and Wcxley, 1981; Thorndike, 1949).
Comprriron of peer nomiartion, ranking,rnd ratings
by Love (19811 yielded reliable and valid results. No birr
on recount of friendship was observed in these cases.
Bared on r study of the rccurrcy of peer ratings
of British managerr, Imrda (1982) concluded that behaviour
brred informrtion often reducer halo error as the rater is
able to perceive the multi-dimensionality of performance
criteria. Krne and Lawler (1980) have also observed that
peer ratings rre good eerrures of performance.
Humford (1983) observed that multiple raters
counter balance observational errors and increase
rcllablllty. Peers have realistic contacts amongst each
other. Hence, lnsplte of lack of objective measures for
most Larks, peers arrive at valid ratings since they observe
and corprre the task-relevant abilities of their colleagues.
Mreover these data are not significantly influenced by
factor8 such as friendship, race or sex discriminations.
Attention Lr based on performance alone.
Information about performance in a leadership or
arnagerlal role can be useful for peer raters belonging to a
sample constituting different behavioural dispositions. The
idel hrr received conrlderable support from subsequent
rem~atch.rr (Korrrn, 1968; Kane and Lauler ,1978). In the
industrial setting too, they have proved to be effective
predictors (Hayfield g g., 1980; Roadman 1960; Weitz
1958) .
Hayfield g c. (1980) requested three thousand
agents from three different Life Insurance Companies to rate
each other on work orlented and socially oriented items.
Subsequently, 150 agents were promoted as assistant
managers, without considering peer ratings. Supervisors
r a t d them six and eighteen months after promotion. ?he
test retest reliability after a year's interval was 0.77.
Asslstant managers w r e expressed as letter grades ranging
fro. A to E. The chi-squared value representing the peer
rrtlng scores and the percentage of assistant managers who
were rated ruccessful at each letter grade was 12.50
IplO.OS). Further, peer ratings made in one group was not
valid if the lndlviduals promoted were subsequently
trrnsfcrtsd to other groups. Peer ratings were made Ln
crtabllrhed groups where new agents were found to be
nornally included only on the termination of established
88tntr. Uhere agents are subject to varying degrees of
erpQsure, i t 11 porrlble chat new agents might receive low
ratfnp as thelr rbllities have not been demonstrated as
derkrd. Bared on length of service data Mayf ield 9 g. (1980) rho& that peer ratlngs sre unlikely to be effective
predictors for Lndividuals with less than six months
experience.
Siege1 (1982) obtained prired comparlson
evrlurtions by prospective peers (N-20) and superiors (N-4)
for r group of twenty savings and loan association branch
managers. Only sl~teen of the peers however provided their
rvaluatione. The superiors were infor~ed that their ratingr
together with peer ratings would be used for the four
promotional declslons they were to make in the near future.
The branch srnagers knew that appointments to higher posts
were to be made shortly. Hovever they did not know that
they were actually participating in the selection process,
Each lndlvldual evaluator's comparison yielded a score for
every branch manager who was being evaluated.
A month prior to the paired comparison, nineteen
of the partlclpants w r e administered a battery consisting
of the 7'h:~rstnne Test of Mental Alertness (W1, planning,
fudgemcnc,cowyrehcnsion and expression tests beLongin8 to
the Flanrgan Indusrrial Tests (FIT), the Strong-Campbell
lntcresc Inventory (SCII 1 , the Survey of Interpersonal
Vatues (SIV). the Supervisory Index (SI) and the Cuilford-
tlmserran Temperrment Survey (CZ).
2-score conversions of assessments made by paera
rml aup.riorr facil itated selection of three incumbents rr
the two retr of evaluators seemed to agree alike. In the
fourth instance however ratings by peers far exceeded those
of the superiors. Since the candidate was found more
favourable to the peers who would subsequently become
rubordinates, the superiors accepted the candidate for
promotion,
Correlation betueen the mean effectiveness scores
by peers and supervlrorr was 0.81. Interrater correlations
ranged ktueen 0.31 and 0.86 in care of peer comparison
matrlx. Interjudge agreement was found to be high,
excendlng to ocher independent prychologlcal characteristics
of the subjects evaluated. The peer evaluations enabled
rcccptrblc promotional decisions in this study. Discussion
by eupervlsors, of the differences in the evaluations made
by them indlvidurlly revealed Important differences in
percept tons about the determinants of managerial
effsctivcness. Two important views emerged from the st*.
One attached greater importance to human relations skill,
*l ie the other tended to emphasize knowledge about
financial matters as s more important component of branch
manager effectivences. These two aspects were considered
to k complomrntary to each other by the group of superloro.
The purpose of the study conducted by b u n t (1981)
vclr to compare the psychometric properties of subordinate
ratings of managerial performance to those of supervisor and
@elf ratlngs. Eighty middle level managers, their immediate
superiors (N-80) and their subordinates (N-3651, from a
hightechnology multinational corporation participated in the
study.
A questlonnalre was administered to the subjects
to arress managerial effectiveness. The first section
consisted of items pertalnlng to background information of
the respondents. The second contained eight behavioural
items Iknov-how, adcinlstration, training, work direction,
feedback, motivation, tnnovatlon and consideration).
The multitralt-aultlmethod analysis was used to
study convergent validity, discrininant validity, leniency
error a n d halo error. Coaputation of Duncan's Multiple
Range Tcst revealed that self ratings were the most lenient
(H-6.471, folloved by supervisor ratings 1H-6.27) and
subordlnatr ratings 1Mw6.02). Differences in mean levels of
ratlngr war si~nlficanc across rating sources in case of
know-how, vark direction, feedback, motivation and
Data shoved evidence of convergent validity and
halo affect. Discrimination asong trait8 was observed to be
large. Results of ANOVA confirmed that differences anon&
managers are a t t r i b u t a b l e t o t h e r a t i n g sources .
Discriminant v a l i d i t y was higher than convergent v a l i d i t y
but equal t o halo e f f e c t . The highest convergent v a l i d i t y
was betueen super io rs and subordinates ( 0 . 2 4 ) and t h e lowest
(0.16) between super io rs and s e l f ra t ings .
I n t e r r r t e r agreement was assessed through
I n t e r c l a s s c o r r e l a t i o n s . Agreement was g r e a t e r i n c a s e of
(Training rnd Hotivat ion) than on (Delegat ion and
Cons idera t ion) .
Comparison of r e s u l t s obtained i n t h i s s tudy v i t h
chose of o thers revealed t h a t t h i s study yiefded smal le r
convergent v r l t d t t i e s and ha lo e f f e c t and g r e a t e r
dlrcr lminanc v a l i d l c y . But the same could not be s a i d about
s e l f ra t ings . The r e s u l t s a l s o i n d i c a t e t h a t i n t e r r a t e r
agreenent I s g r e a t e r on some dimensions l i k e t r a i n i n g and
e o t i v r t i o n than on o t h e r s such a s de lega t ion and
cone ldcrs t ion . The study i s perhaps one of t h e f i r s t t o
analyze psychometric p roper t i es of subordinate r a t i n g s .
Since rubordlnarc r a t lngs a r e p o t e n t i a l sources of
performance information i t s use has been recommended. Use
of a t l e a s t th ree subordlnates ' r a t i n g s f o r a s i n g l e manager
has bean suggested.
The f ind ings imply t h a t t h e use of mul t ip le r a t i n g
sopltces would ensure inc lus ion of more usefu l informreion i n
appraisals. Inspite of possibility of low interrater
agreement its use is emphasized because of greater
predictability. Since some dimensions are rated more
reliably by some rating sources, it is inevitable that they
be included in the appraisal process.
Heyers (1980) observed that self ratings vere
unrealistically high. Thornton (1980) dissuaded its use for
ad*inistrative decision making. Bauman and Dent (1982) on
the other hand observed that the perception that self
assessments would be compared vith other objective measures
1ncreas.d the accuracy of self ratings. The viev was
supported by Mabe and West 11982).
The use of self assessments for performance
evaluation was recornsended by Levine 11980). Locke and
Lathre 11984) demonstrated its effectiveness in their goal
rettlng theorv.
Caldwcll and O'Reilly (1982) studied individuals
in r large franchise organization. The study demonstrated
the lmpllca~tons of self-monitoring, The individuals vere
found t o perform better If they vere sensitive to social
cue* and were able to modify their behaviour to situational
rcqulrrmrnts.
fsul and Ohlott, (1988) studied three sources of
disrgreem~nt in the effectiveness ratings via., rating
errors, selective perceptions and variations in criteria
weight. A11 three causes provide reasons of varying degrees
for the low agreement coefficient. Three hundred and forty
four managers of a Fortune 100 company participated in the
study. The ratee managers were given,six questionnaires and
requested to complete a questionnaire and distribute one
each to a superior, one peer and one subordinate vith whom
the ratee wrked least well, and one subordinate and peer
vith whom the ratee interacted frequently. Confidentiality
was assured.
Three hypotheses on criterion type and criterion
weight found consensus in the effectiveness ratings of
uuperiorr, subordinates and peers. Agreenent was strongest
betvcen superiors and the peers. Agreement among the raters
was high on role behaviours and personality traits of the
managers as criteria for effectiveness. Role behaviours
accounted for largest variance in all three models ( R ~ - 0.07, 0.11 and 0.12) for superiors, subordinates and peers
respectively. P-values indicated that the criteria for
eEfectlveners did not significantly differ among the three
sets of raters. F-statistic was not significant on either
the personal or the role behavlour variables. Raters at the
tame level tended to assign similar criterion weights.
The results suggest that rating error tendencier,
informational differences and criteria differences are
possible causes for low interrater agreement. The study
emphasirer the need to consider view points of different
rating sources to arrive at accurate explanations of
managerial effectiveness.
In yet another predictive study, (Kraunt, 1975)
showed that peer ratings taken under a "real" set could be
used among middle and higher level managers to predict later
success. Peer ratings were obtained from 156 middle level
aanrgers and 83 higher level executives. Two years later it
was found that It did not predict performance of the middle
eanager sample, however 7 of the 13 rating dimensions
prcdlcted promotlons significantly (median r-0.35). The
highest predictive validity coefficient vas shown by
mobilltv (r-0.U). On the other hand for the executive
grade ten of the peer ratings significantly predicted
promotlons. The highest validity coefficient in this group
was general Impressions (r = 0.44). According to Kraunt
( 1 9 7 5 ) , the use of peer ratings for evaluation of managers
may merely foraalize a practice in which managers are
a1 ready invoivcd.
Results of a study on "feedback-seeking"
behaviour of three hundred and eighty seven managers, as
observed by their superiors, peers, and subordinates are
reported by Ashford and Tsui (1991). It showed that
managers' inclination to seek negative feedback, increased
their understanding of the evaluation techniques adopted by
these feedback sources. Uhile seeking negative feedback
tended to positively influence the opinions of all the three
categories of raters. As regards overall effectiveness,
seeking positive feedback decreased the same.
The authors opine that the willingness of
effective managers to seek negative feedback may be a result
of their confidence. Similarly strong egos might have
prompted them to use direct enquiry while seeking feedback.
Velsor g . (1993) examined the relationship
raong self-perception accuracy, self-awareness, gender, and
leader effectiveness. Conclusions were based on data
provided by three groups of managers 1N-668; N-168 and
N-79). Results show that discrepancy is a result of
diifcrer~ces between self ratings and ratings by others. The
finding vas consistent across all three samples, and no
gender differences were found. Underraters constantly rated
themselves lover than the accurate or overraters. They were
ratmi highest by their direct reports. Overraters are rated
lowst by their subordinates. However they tended to
consrantly rate themselves high. The data also suggested
that underraters are the most effective eanagers. They were
alro found to be rated highest in self-awarenesr based on
direct reports. Overraterr were perceived as the lowest of
the three groups In self awareness and effectiveness.
Nilsen and Campbell (1993) studied the differences
between the perceptions of subjects and their raters.
Discrepancies In perception were observed on the 360 - degree feedback instrument. The discrepancy on skill based
multiple rating instruments related to dlscrepancy on
personalltv based multirater instruments. Self perception
Is contended to be a stable indlvidual difference. Self-
other dlscrepancy was also found to be stable over time.
A salient feature of studies reported here Is
thelr emphasis on self, peer, and subordinate ratings. A
norevorthy point in peer assessments is that researchers
have shlfted their attention from stressing reliability and
valldlty to the study of reactlons of actual users besides
attempting a theoretical understanding of the reasons behind
hlgh rellablllty and validity (Latham, 1986) . The concept
of self assessaent as a complementary aspect of formal
appraisals is also gaining mosentum. Organizations must
replace formal appraisal by superiors alone with multirater
cachnlques so as to ensure more accurate evaluations of
sanagerirl prforaance.