Selective Perception

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 23

* Academy of Management Journal

1997, Vol. 40, No. 3, 716-737.

THE SELECTIVE PERCEPTION OF


MANAGERS REVISITED
JANICE M. BEYER
University of Texas at Austin
PRITHVIRAJ CHATTOPADHYAY
ELIZABETH GEORGE
Western Michigan University
WILLIAM H. GLICK
Arizona State University
dt ogilvie
Rutgers University
DULCE PUGLIESE
University of Texas at Austin

Results of a systematic replication of studies by Dearborn and Simon


and by Walsb suggest tbat differences in tbeir conclusions resulted from
differences in tbeir experimental procedures. Managers are broader
perceivers when they are encouraged to identify more problems. Also,
contrary to Walsb's and otbers' tbeories, in tbis replication belief struc-
ture did not mediate tbe relationship between functional experience
and selective perception. In addition, predominantly negative relation-
ships were found between areas of experience and perceptions, indi-
cating tbat functional experience may produce selective imperception
as well as selective perception.

Organizational researchers agree tbat many important decisions taken in


organizations involve complex and ill-structured problems likely to be af-
fected by managers' personal cognitive processes (Hambrick & Mason, 1984;
Mintzberg, Raisingbani, & Tbeoret, 1976; Starbuck & Milliken, 1988). An
early, widely cited empirical study showed managers' "departmental iden-
tification" was related to tbe types of problems they identified in a case
(Dearborn & Simon, 1958: 140). Tbis finding bas been interpreted as evi-
dence that functional experience selectively cbannels tbe perceptions of
managers when tbey solve complex problems.
Walsh (1988) replicated and extended the Dearborn and Simon study.
The authors wish to thank Kathleen Knoll, Smita Saxena, and Richard Trezza for their
research assistance and the Bonham Fund of the University of Texas at Austin for its support.
The authors also wish to thank Vicki Goodwin, James Walsh, and Kimherly Williams for as-
sistance with the instruments and measures used in this study, and Sim Sitkin and James Walsh
for their comments. Authors are listed in alphahetical order.

716
1997 Beyer etal. 717

postulating that functional experience was incorporated in managers' belief


structures, which served as a cognitive filter that narrowed their perceptions
toward related information. He found that (1) the content of managers' belief
structures was often broader than their functional experiences, but (2) there
were only a few scattered relationships between the content of their belief
structures and their information processing.
Although the results of these studies have been interpreted as refiecting
"selective perception," these studies actually measured this construct in a
special, limited sense. Both sets of researchers operationally defined selec-
tive perception only in terms of its direction. They assumed that selective
perception is manifested in information processing directed at content that
matches current functional role or functional experience because familiar
content is most meaningful to the perceiver (Lord, 1985). They did not focus
on the actual selectivity of the perception—that is, the breadtb or narrow-
ness of subjects' information processing. This limitation is especially curi-
ous in the second study, since Walsh did compute a measure of the breadth
of his subjects' functional experience. The same logic should be applied to
perceptions that he applied to functional experience. Clearly, the most basic
concern in measuring managers' selective perception should be whether the
managers are general, broad perceivers or specialized, narrow perceivers.
The issue of where they direct their perceptions is logically secondary.
The study reported here was undertaken as a systematic replication
(Aronson, Ellsworth, Carlsmith, & Conzales, 1990) and extension of the
Dearborn and Simon (1958) and Walsh (1988) studies. As Aronson and col-
leagues defined it, a systematic replication is "a repetition of an experiment
in which the experimenter systematically varies some aspect of the original
conditions, procedures, or measures in order to resolve ambiguities or add
new information . . . and provide for a clarification or extension of original
findings" (1990: 351). They argued that systematic replications "are invalu-
able for uncovering the source of inconsistent results and can ultimately en-
hance our understanding of conceptual variables" (Aronson et al., 1990: 55).
These quotations express the purposes of the present study. The meth-
ods of the two replicated studies differed in one crucial respect: Dearborn
and Simon instructed respondents to identify the most important problem in
a case; Walsh instructed respondents to identify as many problems as they
considered important. Because the first method gave respondents a more
limited observational goal (Dweck & Leggett, 1988) than the second, it is not
surprising that managers responded as more selective perceivers in the first
study than in the second. In addition, Walsh's study included a possible
confounding factor—priming effects (Fiske, 1993). By asking subjects to de-
scribe their belief structures before analyzing the case, Walsh made their
beliefs salient and thus likely to affect information use and problem identi-
fication. Finally, as was already pointed out, neither study measinred selec-
tive perception in the most basic sense of whether subjects' perceptions were
relatively broad or narrow.
This experiment was designed to assess the impact that experimental
718 Academy of Management Journal June

methods had on the results and conclusions (Latham, Erez, & Locke, 1988) of
two important studies. It also makes two other contrihutions: (1) we inves-
tigated the breadth, as well as the direction, of perceptions and (2) we un-
covered evidence that people may exhibit selective imperception, in which
functional experience tends to direct attention away from stimuli in some
unrelated areas, rather than toward stimuli in areas related to their func-
tional experience.

MODEL AND HYPOTHESES

To clarify the underlying conceptual differences between the two rep-


licated studies and our study, we developed a causal model for each. All are
presented in Figure 1.
Dearborn and Simon tested only the direct relationships between de-
partmental affiliation^ (expressed largely in functional terms) and selective
perception (measured by areas of problems identified), as shown in Figure
la. Their conclusions, however, mentioned company-wide or departmental
"viewpoint" as a possible unmeasured mediating variable (1958: 142-143).
Following this lead, Walsh (1988) developed and tested the model given
in Figure Ib. He postulated that past experience shapes belief structures that
in turn channel selective perception. However, he reported statistical tests of
only the relationship between belief structures and perceptions.
To be consistent with the aims of systematic replication, we developed
a more comprehensive model (Figure lc). Our model includes the two direct
relationships central in the theorizing of the two earlier studies. In addition,
it includes interactions that represent differences in the experimental pro-
cedures in the two studies, testing (1) whether observational goals (repre-
senting differences in instructions) interact with Dearborn and Simon's sole
predictor variable (functional experience) in predicting selective perception
and (2) whether this variable and salience (representing possible priming)
interact with Walsh's primary predictor (belief structure) in predicting se-
lective perception.
We present further justifications for each of the relationships in the
model in discussions appearing before or after each hypothesis. Paths in the
model are numbered to correspond to the hypotheses.
The two central direct paths in the models used in the two prior studies
predict the following:
Hypothesis la. The greater the experience of decision
makers in a functional area, the more that functional area
will be represented in their belief structures.
Hypothesis Ib. The greater the experience of decision

^ Although Dearborn and Simon measured current affiliation, their discussion of "selective
perception as a learned response stemming from some past history" (1958: 140) indicates that
they expected this measure to include experience as well as current status.
1997 Beyer et al. 719

FIGURE 1
A Comparison of Models Employed in the Two Earlier
Studies and the Present Study

Functional Affiliation/ 1 Departmental Selective Perception


Experience I Viewpoint
_ J
(a) Implicit Dearborn & Simon (1958) Model

Functional Experience ...y, Belief L ^ Selective Perception


' Structure •
1 J
(b) Walsb (1988) Model

Salience
5
>f
Functional Belief Selective
J<
Experience Structure Perception
i
Observational
Goals 3*

lb '^i
(c) Model of Tbis Study

Relationship or concept not tested statistically.


Relationship tested statistically.
* Relationship supported by our results.

makers in a functional area, the more that functional area


will be represented in their perceptions.
Because they are intended to replicate past findings, as already discussed,
both of these predictions clearly refer to the direction of perceptions. The
purpose of Hypothesis la is to address a gap in Walsh's analyses. Hypothesis
lh is intended to replicate the major finding of Dearhorn and Simon's study.
Walsh measured subjects' perceptions by the types of information they
sought and used and the problems they identified in analyzing a case. He
found only marginal support for the relationship between belief structures
and problems identified and scattered relationships between belief struc-
720 Academy of Management Journal June

tures and information used and sought. Many researchers, however, have
theorized that people use schemata, or cognitive maps, to simplify complex
stimuli and thus channel their perceptions (DeNisi, Cafferty, & Meglino,
1984; Kiesler & SprouU, 1982; Prahalad & Bettis, 1986; Weick, 1979]. Since
null or weak results are not definitive, we used the following hypothesis to
again test these ideas concerning the direction of perceptions:
Hypothesis 2. The more decision makers' belief structures
emphasize a functional area, the more that functional
area is likely to be represented in their perceptions.
As already mentioned, a major difference hetween the Dearhorn and
Simon and the Walsh studies was the instructions given to subjects. Instruc-
tions by experimenters can set observational goals (Devine, Monteith, Zuw-
erink, & Elliot, 1991; Neuherg, 1989; Pavelchek, 1989]. Dweck and Leggett
suggested that different goals stimulate different response patterns of cogni-
tion, affect, and behavior, leading decision makers to approach ill-structured
situations with "different concerns, asking different questions, and seeking
different information" (1988: 260]. It follows that a goal of identifying a
single, important issue will evoke different, probably simpler and more re-
stricted, information processing than will a goal of identifying multiple
prohlem issues:
Hypothesis 3. The more restricted the observational goals
of decision makers, the more selective their perceptions
will be.
Unlike our other hypotheses, this hypothesis predicts selective perception
in terms of the narrowness or breadth of perceptions.
Also, hy encouraging more filtering, the goal of identifying a single issue
may amplify the effects of experience and belief structures on information
seeking and problem identification. One explanation for such an effect
would be that people seek to economize on their cognitive efforts. If asked to
look at one problem, the prohlem that is easiest for them to identify may be
the one most closely related to their functional hackgrounds.
Hypothesis 4a. The more restricted the observational
goals of decision makers, the more likely that their func-
tional experience will be reflected in their perceptions.
Hypothesis 4b. The more restricted the observational
goals of decision makers, the more likely that their belief
structures will be reflected in their perceptions.
These hypotheses again concern the direction of perceptions.
The other experimental factor relevant to our replication was the pos-
sible priming that Walsh gave his subjects by having all of them respond to
his measure of helief structure before they performed the experimental task
that produced measures of their perceptions. Priming occurs when people's
attention is directed to certain stimuli that are relevant to what they are then
expected to do. It has been consistently shown to affect how people perceive
1997 Reyer et al, 721

and process information (Fiske, 1993; Squire, Knowlton, & Musen, 1993).
Similarly, tbe recency witb wbicb people have been reminded of associated
ideas or information (Hoch, 1984; Tversky & Kahneman, 1974) has been
posited to affect tbe retrieval of relevant information and scbemata. More
recent stimuli are more salient. Furthermore, the salience of a particular cue
or set of cues tends to give salience to the categories or set of categories to
which those cues are relevant (Ilgen & Feldman, 1983)—in this case, to
functionally related categories. It thus seems appropriate to conceptualize
and test the issue of priming in the Walsh study in terms of salience:
Hypothesis 5. The more salient decision makers' belief
structures are, the more likely it is that their belief struc-
tures will be reflected in their perceptions.
This hypothesis also concerns the direction of perceptions.
The five sets of hypotheses presented above serve to comprebensively
test tbe implications of tbe arguments advanced and the procedures used in
the two earlier studies. By including interactions as well as direct relation-
ships, we addressed questions that might arise about both indirect and direct
effects of all our hypothesized predictor variables. By including bypotbeses
tbat address two operational definitions of selective perception, we sought to
be comprehensive in investigating the concept itself and in testing the theory
that has developed around it.
Although Dearborn and Simon (1958) suggested tbat "selective atten-
tion to part of a stimulus reflects a deliberate ignoring of the remainder as
irrelevant to the subject's goals and motives" (1958: 140, empbasis ours),
neitber tbey nor Walsb explored this possibility. Although we did not de-
velop hypotheses to test this speculation, we considered it an important
issue and therefore explored it with additional analyses.

METHODS
Procedures
Individual subjects reported to a designated room wbere two or more of
the researchers explained the experiment and passed out packets of experi-
mental materials. After signing a consent form, all subjects received the
following tasks to perform: analyzing a case (the same one Walsh used),
completing a questionnaire-type instrument measuring their belief struc-
tures, and completing a brief questionnaire tbat measured their work expe-
rience and other demographic characteristics and contained a manipulation
check. The sequence of tasks each subject was to perform was indicated by
numbered envelopes that contained the instructions and instruments for
each part of the experiment.
As subjects analyzed tbe case, tbey were told to underline relevant
information and were given one of two sets of instructions reflecting differ-
ences in tbe two earlier studies: to replicate Dearborn and Simon's (1958:
141) instructions, we asked some subjects to "read tbis case with an eye
722 Academy of Management Journal lune

toward identifying tbe most important problem facing the Arbor Company,"
and to replicate Walsh's (1988: 878) instructions, we asked other subjects to
"read this case with an eye toward identifying all of tbe important problems
facing the Arbor Company." Subjects filled out the other instruments in-
cluded in the packet either before or after analyzing the case, according to
the condition to which they had been assigned.
Subjects were thus randomly assigned to one of four conditions on the
basis of both the number of problems they were asked to identify and the
order in which they performed the tasks. The four conditions were: (1) single
problem, case first, (2) single problem, case after measure of belief structure,
(3) multiple problems, case first, and (4) multiple problems, case after mea-
sure of belief structure. Subjects were debriefed on the true nature of the
experiment after all had completed their tasks; we then paid them and asked
them not to discuss the experiment with others.
Subjects
The two earlier studies used executive students as subjects. Subjects
participating in this study were 137 regular full-time master's of business
administration (M.B.A.) students recruited at a large southwestern univer-
sity; 120 of them provided complete data. Each was offered ten dollars for
participation. None were currently enrolled in the researchers' classes. They
were told that they would be participating in an experiment in managerial
decision making and assured of the confidentiality of their responses.
The subjects had the following demographic characteristics: 87 were
men (72.5%), and 33 were women (27.5%), Their average age was 28.06
years. They had an average of 4.71 years of work experience. Thirty-four
(28.3%) expected their first job after finishing the M.B.A. program to be in
finance or accounting; 33 (27.5%) expected it to be in marketing or sales; 1
(0.81%), in personnel and human resources; 9 (7.5%), in production and
operations; 8 (6.7%), in information systems; 3 (2.5%), in R&D; and 23
(19.2%), in general management. Seventeen subjects (14.2%) were foreign
nationals.
Measurements
Functional experience. We asked subjects how many years they had
worked and then asked them to divide their total work experience by per-
centages into management areas, which we collapsed into seven categories:
finance and accounting, human resources, production and operations, in-
formation systems, marketing and sales, research and development, and gen-
eral management. The measiures of specific functional experience used in
our analyses were the percentages of work experience in each of these cat-
egories.
In his replication, Walsh (1988) used a global measure of functional
diversity that assessed the extent to which managers had varied functional
experience rather than experience in a single functional area. We did not use
this measiu-e to test hypotheses regarding the direction of perceptions be-
1997 Beyer et al. 723

cause the theory suggested that experience in specific functional areas


would be related to beliefs and perceptions in the same or related functional
areas.
Walsh's measure did, however, seem suitable for testing our hypotheses
about the breadth of selective perception. Consequently, we also constructed
a global measure of the breadth of functional experience that assessed the
extent to which respondents had varied functional experience or functional
experience in a single area. The measure, based on Walsh's (1988) formula
for his functional diversity measure, is as follows:

Breadth =
where x,- is the number of years worked in a particular functional area and X
is the total number of years of work experience.
Belief structure. Because Walsh (1988) found that only 2 out of 15 re-
lationships were marginally significant when analyzed with his measure of
belief structure, we decided to try a different way of measuring them. We
modified and further developed somewhat simpler measures (Glick, Miller,
& Huber, 1993) that assessed (1) the importance executives assign to 17
operative goals suggested by the work of Quinn and Rohrbaugh (1983) and
(2) their cause-effect beliefs regarding the efficacy of each of 22 business
tactics taken from the work of Porter (1980) and of Robinson and Pearce
(1988). Because this measure was not specifically designed to reflect func-
tionally related belief structures, we extended the initial 30 items to include
an additional 28 items^ focused on functionally related beliefs. Respondents
were asked either (1) "In your opinion, how important is it for business firms
to maximize ?" or (2) "How important is it for managers of business firms
to pay attention to the following ?"; all responses were on seven-point
Likert scales anchored hy 1, "not very important," and 7, "very important."
Together, the new and old items measured the same functional areas we
measured for functional experience.
We factor-analyzed these items to see if the data supported our a priori
expectation that belief structures are organized along functional lines. We
used a common factor model with oblique rotation (Mulaik, 1972) because
we expected that dimensions of belief structures would be correlated. A
six-factor solution with 27 items produced distinct factors for finance and
accounting, human resources, production and operations, information sys-
tems, marketing and sales, and research and development. The interitem
reliabilities (Cronbach alphas) for scales based on these factors ranged from
.64 to .83. We used these scales in hypotheses testing the direction of selec-
tive perception.
Observational goals. Observational goals were manipulated in the case

^ These items were generated by a group of researchers with different functional back-
grounds following observations of ten groups of M.B.A. students discussing an integrative
strategic management case.
; i n i i l \ .-^is | I ; K k u L O n t i n ; i i i \ " i ' r ' .si)i!i:t I n r t h n r a s r , i n ( i o n t h i ' n i s j K H i s i ' l o r n i s .
^ u h i f ^ ' N w f ' i ' i ' i n ^ l r i i r t f ' d t n i d i ' i i i i t X f i t h p r "tbi.' i i i o s i i m p o r i n n t p r o h l c n i " i n
i h t ' 1 <i'^i! n i • • ( ; / / n\ t l u ' i n i j ) i i i i . i i ] t p r n i i l i i i i i s " i n l l n - i . a M ' ,
T l i r l;i^i c n v i i l u f i c I'm <]l! s i i h i e r l ^ r o i i l . i i i H ' f i a i n a n i j i u l , i l i n n i i n M . k , S u h -
!•'( t-^ W i i [ ' ( ' l i ' ^ k M t l l l u ' r \ l f ! l 1 I n w h i c h \'h''-\ i i g r ' i ' i ; i l SM ([i,-,;n^i I T I I w l t l l t l i r
! l l l l l l ^ v ! I l , ^ l h r ( H ' s t i i l r i i i i i n l s (1) " i vv,is o i i K ^ i f x i k i i i i ; I n r i h r n i n s t i n i p o r L i i i t
p r o h l e m . i i c a . " (^1 " I n n t i M M i n i a i u i i i n r r | . H - n h l r i n s t l u i n 1 w a s a s k e d I n
i i i i u i l i l \ .*' a n H ( : i | ' T h r i n ^ t i ' m l i u i i s i . i n i s t r , l i i i c i i \\\r. t h n n n i n h i i r s s ni m y i : a s n
a n a U s I s , " Sii_;ni!'i( a i i t i l i l h ' i ' ' ' m , r s w r r t ' n i j s n r x ' i ' J h r i w c c n \}i>-- [wn nxpt'ri-
m f ' l l t . i l I o m i i i i o n s I n r r , n . h n l [ l i t ! t h r n ^ i i r i i i s (/ L!..">. /' < . 0 1 : / 7 . 1 4 . p •='
. n O O l : / - 4 . I I ) . /' • . ( K I O l 1, A s n i l a H d i t i o i i a l \ v a \ n f r h r c k i i i L ; i h r m a i l I p u l a -
l i n n , \\i' a l s f t I ' S . i m i n c i l w l i r t h i ' f s u b j e r N w l m w i ' i ' r a s k n l t u H l f n l l t \ t h r
n i o s i i l l i p i i r l a n l j i i ' n b i i - i i i r c p n r l i u l i n w i u ' [ n ' o h l n m s t h a n s n l i i r c t ^ w l m wv-vr.
a-^krd i n i ( l r n l i r \ lill l l i r i m p n r i a i i l p i n h l r n i s . T i n ' anal\-si,> - ^ l u j w ' d thri! s u h -
j i i c l s i n t h r h i i ] ; j , i i ' - p i - n b | [ ! i n i l u i d i l i n i i i d i i i i i i l ' i t n l s i g i i i b i a n l l \ (/ - A.'.VI, p •
, 1 ) 0 1 ) I r w t M p r n b l c i i H i b a n s i i b j r i : ! - . i n \hv n m i t i p l c - p n i h i c m > ' t i i d i t i u t i .
S i i l i e n i . i ! o f b e l i e f ' s . S i i l l c m i- n t h c l i i d s w a s i i h i i i i p i i l . i l i ' d h \ i h f { ( I ' d c i i d
w i i i i . h s i j b j i ' ( Is \\vv\- i i i s l r n c l c d l o p c r t o n i i l a s k s , T b n s i ^ i n j h c b d l i c l s - . s a b c n t
t n r i d i i i o n l i l l i M l u n t t i H ' b i d j i d ' s i i ' i u \v\\v i u s l r i i i M f n t b c l n i f a i i a K / i n ; ^ t h e ) A'^r\
l l m i i t h r r s i i b j o c i s l i l l i u i il o u t a l l r i ' . i i i a U ' / j i i g I l i r ( .isr
S i ' l e c t i v e p t M T H p h o T i . T w o i n d i i dtca's o i ^clfM ti\"(' p r n r p l i m i w c i * ' u s r d
i n our- a i i a l v s c s : i n l o r n l a l i o n a t t i ' i i d r d l o <ind p r n b l t ' i i i s i d i u i l d i r i k ' l i n t h Wiir'n
I n d c d . ai I n r d i n ^ u t(] i i i s t r u r t i n n ' ^ n b l i i i i u n i I r u i n W a l s b f ] 'I'Hi! p ^ ] s m u ] t l u n -
a l i o n j , i n i o \\\v s a m e l i \ ' i ' I ' u n i J i n n a l a r r a s h e I W I M I : l i n a i H f , i n d a i -
n^. linnian mlalions, lAlornal manaiirniiMit. m a r k r i i n u , and i n l r m a l
n i a i i a n r m i ' n t A h h o n u b bis r a l r ^ o r i r s d o not cxai ll\' mat: b t b r sc\-cn l u m -
l i o n a ! a r r a s o l r x p r r i m i . n \\v n u i a s u r i M l , i b r y a i ' r ( | n i l r r l o s r . W e n s r d b i s
i a h > ; _ ; n i i p ^ i i i i i a U N r v\"t: w a n t e d t o r i ' p M r a t r b i - - n i r a ^ u r r s o j p i a ' c r p H o n ^ a s
i , j o s r l \ LIS p o . s s i b l r .
T w o o | i l i f i r s i M u brr>. ( l a i i i f d ^ o o p r r a t i v f K h i n s r h i s i la'^siTu , i t i o n
> r b r m r ( W a l s h . I ' t o B ; V>'''\] n n l i t t b r \ ' b a d r r a i J i r d ai i r p t a i i l r h^x^rks o i r r i i -
a f n l i t \ - S i i l ) s r i [ u r n l to t h i s trainin^..^, i l u ' t w o i^oilri s i n d r p c L i i l r n I l y r o d i - d t h r
I n i n p h ' t r M i o t t l i s r s I n r i h r i n d i ^ a i n r s t i j . s r l n l i x ' r j i t ' r c r j it i o n , i n t r r r a h - r
i T l i a b i I i S h " ^ , c a l c n K i U ' d a s i t i l i ' a c i a . s s c o r r r l a i i n n i o r l l i c i r n i s {M'X'. U^kb
S b i o n t \- [ ' ' I r i s - ^ . l ' ) 7 M k w f M ' r , i s f o l l u w s ' t o r i n i o i a n a l i n n s n n g h t , . T o . t o i '
i n t o r i n a i i n n a l t r i n h ^ d t o . , 9 8 ; a n d ior' p i ' o b l i t i i i s i d i i n l i t i r d . . 0 1 ,
A*^ ( i i s r n s s c d t M r l i r r . s r l r c t t v r p ( ' ? ' L r p t i ( j n ( a i l I H > ( o n r r [ ) l n a l i / ( ! f l i n
l i a i n s o i r i d i r ! ' (1) till' ;_;rn(!ral n a r r o w n e s s o i h [ r , i i l t h n [ pta'i r p t i o n s o r \ 1 )
w b r l h r i ' p r r r r p t i o n . ^ iirr d i r r r l r d t o w a r d i n i o j ' n i a l i n n j'idalcd i o inns l i n n a l
' • x p i ' t a i M K e . O n r a n a l x M ' s m i p h n i ' d J o n i n i r . t s u r e s o t i i i r n a n n v . ' ; ) ! •-.-•
luvi.u.Uh of j H ' r n n p l i o n s : I h r l o l a l a i n o n j i l oi inlnrniaSiini sui)|r(;ls a t t r i H i c d In
(i,f'., t h r n n i n b i M ' o f ^ i * u ! n r n t s o t t r x t n n d i a l i n r d ) . t b r m n n h r i ' (jf b m i t i n r i a l

Isii M i l - .\ .u ; ! ! t h w r i t v. i^ij
iii-;jiiri,iiinn I M llii- -.1 n l v .
ar'eas in w b i o l ! lliev a t t e n d e d In i n l n r n i a t i o n , I h e total i n n n h e r nl p r n h i e n i s
llii;\ i d i i n l i l i e d (i.e.. t b e t n n n b e r nt' d i i ' t e r c n t p r n h i e n i s i d e i i t i i l i ' i l ) . a n d d i e
i i n i i i b r i ot h i n t l i i m a l a r e a s in wliii b thf'\ i t i r n t i t i e d [lrntileiiis l\'e .dsij j i s e d
t e n i n e a M U ' e s of t h e l i i r i u ' H n n ui pei'i e p t i n n s , c n i n i t i n L ; b n w [nin b i n t n r n i a -
t i o n snbjiirt.s a t t e n d e d tn a n d hnvv n i a n \ ' p r n t i l e i n s tbe\' i d e n t i f i e d in eat'b ol
t b e l i \ ' e l n n r : i i ( ! n , i l a r e a s n - - e d h y W ' a N h . W ' e t h e n u:^['i\ I n i ^ a r i t h n n i t r a n s t n r -
n i a t i n n N nt t b e n n n i h t ^ - ot p r n b l e n i s in e a i J i a r e a arid ul t b e t n t a l n u m b e r n!
f i i ' o h l e i n s iii ill] a r e , I S t o a d j u s t l o r t b e s k e w n e s s n l t l i e d a l a a m i I n I ' e d l i c e t b e
ellecl nt exti'iMne seni'es tai llltise \ a f i a h l e s .

Analyst's

I ' a r l f . a n n n i e a l ( . o r r e i a l i n n s Y ^ - e r e u s e d a.s i n i t i a l n n i n i h u s t e . s t s l o r b v -
p n t b e s e s w i l b n n i l t i p l e i n d e p e n d e n t a n d t l e i j e n d e n t \ ' a i i a b ! e s (1 l v i H ) t h e s e >
l a . l b . 2 . 4 a . - i b . a n t i .")] t l e . i l i n g u i t b t l i o d i c e i H f i ! ! n l p e r f e p l i n n s , I V i n ^
r e g r e s s i o n a n a l y s i ' s Io test all ot t h e s e h y p n t h i ' s e ^ w o u l d h a \ ' e inflatt^ti t h e
e x p ( ' i - h n e n t - v \ i d e er'rnr I'atc h e \ n n d a r o e p t a h l e l i m i t s boi i i x a n i p l e . t h e ( M -
n n i i i t ,d e n r r e l a t i o n a n a h / s i s ior' i n t f H ' m a t i o n aMi']ultMJ to m ] I n n s iitHiril e x p e -
r i e i ) ! e iiti b i d e d > e \ ' e n [ n e a - - i n ' e s o t t n n t t i o n a l e x | H M " i e n i e a n d t h ' e j n e a s u r e s
oi i i d o i a f i a l i n n < i M e n d e d In in s p e i . i l i e I n n o t i n n a l .ii'eas. T h r a n a U ' s e s nl in-
l e r a i J i o n s ol Cihserv'atiojial j.^uals w i t h t n n i l i n n a ! e x p e r i e n c e i n r l n d i ' t i sex e n
ifltcrai tion.s ot h u K t i n n a l e x p e r j e j u e w i l b ob--ei \ a t i n n a i uo,iU o n t h e . s a m e
l i \ e i n e a s n j - e s t n r i n t n r n i a l i o n a t t e r n t e d t o . \'\ p o ) \ a d i n ; . ; n i i i n i b u ^ l e s t s . t b e
i:annini;al eoia-elal i n n s h e l p e d u s identiK- sels of i n d e p e n d e n t a n d d i ' p e n d e n t
\ ' a i ' i a h ! e s w i i r i s e r e l a t i n n ' - b i f j s m e r i t e d l u i i l i e r i n \ e-.ti^_^at i n n w i U i r e ^ r e s s i t H i
anal\ ses.
W e t b r n u s e d r e n c e s s i n n a n a l y s e s t n t i n i b e r h^st I'el.il i o i s s b i p s t h a t b < i d
p r t ) d u ( e i i si e n i Iii a n t I'elat i n n s h i p s i n t b e p a r t i a n n n i e a l ( o r r e j a l i n n s . T h e s e
r e o j - f s s i o n a n a h se.s t h u s s e r \ e d a s Ib e t e s t s nl t h o s i * b \ - p n t b i w e , s t b a l ( o n i d
nt)t b e I ejei t e d t m Ib e b a s i s o t t h e \ a n t m i c a i l e s u h s . [-'irsl. w e n->ed i ' e ^ i e s s i n n
a n a i v s e s In lest relal i n n s h i p s helvviM-n s p e c i l i t a r e a s nf f u n e l ioTial e.xpi;rientA!
a n d sjH't ilit a r e a s nl' i n f t i r n i a t i o i i a t f e m i p d [n o r s p e i ilii. a r e a s ti| p i ' o b l r i n s
i d e [ ] t i t i e d . T l i e s e < i n i i l \ s e s n i \ e s t iLI,ated r t i s u j l s r e l e \ - a n l I n i b e l i \ p n t l i e > e - ,
etMlccailiili^ i b e d i i ' e e t i o n nl p e n , e p t i o n s . W'r d i d (lie r e n r e s s i o n a n a U ' S e s in
l\\ n \ v a \ ' s . ' H i e fii'st \ \ \ i s t o a n a h ' Z O t h e j - e l a l i n r s s h i p b e t \ \ ' e e n l o g i i a i l v rf^l<iled
a r e a s nl t u n e t i ( i n a t e x [ i e r i e i H e a n d pei'i e p t i o n in o r ( ! e r tn d e t e r n d n e if s p e -
( itii ! n n ( t i o n a l e x p e r i e n c e d i r e c t e d a t l t i n t i o n t n w a r d r e l a l e d a r e a s u \ i n t n r
i n a t i o n . bt n t i i e r r a . s e s . v v h e i e siii:li a n n h \ ' i o n s t t i r i i i s p n n d e n o e d id n o t e x i s l .
w e m a d e j u d g m e n l s i-euardiiie wbic h speciJli fnru litmal experiiMU e related
I o v\-hij b s p e c j l i t a r e a n l p i n b l e n i s i d e n t i j ' i e t ] fir i n l n l ' n h i ^ i o [ ] a l t e n d e t i t o . f'tii
e x a m p l e , \\i» r e i a t e t l t u i u h o i i a l e x p e i a e i i L e i n i n i i r k e H n ; . ; a n d s<iti-s a n d i n
I'esear* b a n d d e \ e l n p m r n t w i t b p i t i h l e n i s I d e n l i l i e d a n d i n f o r m a t i o n at-
t e n d e d t o i n e x t e r n a l m a n a e e i i i e u l . I n o i b e r < i n a l \ --t*s. u e o ' l a t e d l u i u t i o n a l
i A p e r i e r u e in inloi'uiatinn s\stem.-,, e c n e r a ! manri;;^enienl, p i ' o i h n l i o n antl
o p e r a t i i i j i s , a n d t i i i a n o e a r n i a e o o n n t iiit; w i t i i p i ' t i h l e i u s i d e n l i l i e d a r i d j n i t ) ] -
niatioii a t t e n d e d Io in i n t e r n a l i n a n a e n n e n t - Al! nt t h e s e a n a l y s e s r n i i t r n l l e d
[ o r t b e e t l e ( ts nl n b s e r \ ' a t iniiiil t^tjaj^.
S'M ( H i d , w n t ' \ a i n i i n * i l \\ie r i ' l . i t i o n s h i p s [ i c t w f n n n i l iwi'ns of i i i n i t i n n a l
r x p r r i i ' i H >' , i i u ] a i l a r e a s u f i n n b i r n i s i d e n f i i i m l a [if I i n t u i n i a l i o i i , i n p i H l r ( i I d .
\Vi; t h u s I T I . I X I M I t h r a s s u n i p ! i o n that s i ^ l r r l i x T [K^rccptMin tniiri^ tn i n c u s on
only t h n s i ' a r n a s i n \ \ " h i i ; h p c d p l i ' h a \ r h a d r x p i T i c n c n . T l i i > s c i ( i | i'n_i;rcssi(in
c q u a l i d i L s Icl us c x p l i i j ' i ; a n o t h e r p u s ^ i h i l i t x ' — n a n i i ' K . t l i a l hi}u tional c\pc-
riciiM' iliicfteii (r)nniti\-i! [snjccssjnt; a\\a\" IHHII I criain area'-, rathci- ihaii
f i l l u ^ i i i u i( ' i n I e r t a i j i aj'c<i^.

T o t e s t l l v p n i h i i s i s ?,. [ > r j : i l i i : l i n < ^ t i i a t t h e i i . i r i D w i i n s s n r h r o a d t h <'\ p e r -


c e p t i o n s w a l l h i ' a I ' l i i u t i o n u\ o h . s e i ' x ' i i t i o n . i l i;oiii.s, w e u ' - e i l n i u l l i p h i regr'ns-
s i n n a i i a l \ ^ e s . I n I h e s e a n a l \ > e s . w e a l s o M i n t i ' o l l i ' i i t o r t h e e l t r t Is u t \\\m -
t i o n a I e \ p e i i e ] K : e he( a n s e it I'e p r e s e n t e d t h e i o n i p n t in^u e x p l a n a t i o n , ,M I o r cl-
i n g !o t h e o r \ , i o r ^ c l i ' i l i v i i p e r r c | ) t i o n .

KKSULTS
, \ ^ i l l ! I l l a t e d i n i h e M e t h i ul.s ser H o n . w e l e - , t e d H v p o ! [ u " - e s l a . 1 h. :^ 4 a ,
4[i. ,i\\(\ 7\ i n t w o s j a ^ i ' s . [ ' ' i i ^ t , 'A'e u s e i i ( a n o r u * a l i . o i i ' f l a l i u t i (o w e e d ont
i iiMi'l\ i i D i i s i e n i i i c a n l r e l a t i o n s h i p s . SiH'iinit, w e i n r t h e r te'^ied I'cialioiiships
t h a t w i ' i T si!4iii['i( a n i i n I I H ; I a n o n i c a l n ^ s n i t s \,\-ilh m o r e d e l a i l i H l regression
e q u a t i o n s , I ' h i ' s i n n i fit a n t r e s n l l s o l I n i t h o l t h e s e ^ o t ^ o t . m a K ' s e ' ; v\ e r e r a l i u ' r
s u i l t e r e d . U H I t h u s ,ire rcpc.siied o n l \ i n tl)e ti'xt that iollow--,
' i ' h e r e s u l t s o t I I H M a n o n i i a l > o r r e l at i o n a n . i U >es t o j ' i i \ [ i o t h e s c s la and
i h were ,soniewhal n i i x e i l . S p e c i l i c areas o! h n u J i o i i a l expi;['iini(,e w e r e nol
sif^nihianllv I ' e h i l e i l t o ,s]Hici I'ii: h c l i e t s [i i \ polliesi,'^ l a : pa[1 a i l j n s t e d /•/ -:
.1)0), W e i,(included that Hyjidtlic^is la w,is not >np[!orted. This result is
Loii'^i'^Jent with n'Siilts rt'])orted li\ Walsh, Ahhorieh lie d i d not tesl ihi,s
r n l a l i o i i s h i p s l a t i s i ii ( i l l \ \ h i s \ a s u a l h i h j i i ' c l i o n s h o w e d n n n \ i d n n i e t o i ' I h i s
rnialionship in his data.
R e s u l t s o f thf! I a n o n i c a l correlalion a,nal\ses i n d i : ateii, h o w e v e r , that
t\"pe o f h n i c t i o n a i rxperjeiue w a s •-ignil'it a n t U r e l a t e d t o b o t h !\'p>"^ " ^ ^''"•
le( t i \ i> p e r i e p i j c n i - — i n l o r i n a t i o n a t t e n d e d t o ( p a r t a d j u s t e d s ' - ,0^), /> -^ ,(1,1]
and problems ich^nlified ([larl adiu,sled r ' . H I , p < ,()")), ! e n t a t i \ e l y r.up-
portini; H\ pollmsis H i . T o l e s t Ihi.s h \ p o l h i ^ s i s I n l l y , h o w e v e i , u e n e e d e d t o
e m [ ! l i ! \ ' r r ; . ; r e s s i o n a n a h s i s t o n n c o \ - e r t h e ^>[MM:i[ii: r e l a t i o n s h i p s i i n d ( ; ] ' l \ in'4
t h e s e i i \ e i a l l s i g n i l i i iiiit elftli ts.
When we paired snbjcds' a r r a s of Innfiinnal experiem e with the ex-
p e c t e d d i r e i l i o n o l thiM[' p e r c e p l i o n s — t h a i is, w a i l i i o g i i . a l h I'elatod arra.s o l
inlorni<ilion altnnded to or p r o h h a n s i d c n l i h e d ' --wn fonnii no signiticinit

lk

,',U<. ':.>r. a • • \ | j i ' i i " i i ( '• . 1 : n h i i r j , t i ! r > : i •.•. ' . l i i ; ^ - . ^ ^ r i i r r , i t i : i , i i . i , ,i i : , - • ! ! I , [ • ] > i i l i i i ( i n n -


in:: H I I . I I . ' <• . i ! , a ; i i i . M ; M ! . i . ; _ ; " . - i ^ i s r c L i l ' - d I n pniiili'ir.s nh-i;l i!i'-a uiiii ; i i l i •; i i i . i l i n i !
inl'Mihil ]ii,!i',<^jf•nicnl
I ''*i~ /»'i fi r! nl 7'1~

l e h i l i n n s h i p l o r i i d o r n i a l i o n a d e i i d e d t o , VOY p r o h l e m s i d r n t i l i i i d . w e f o u n d
a h ' w s! a l l e r e d s i m i i l i i a n t r n l a l i o n s h i p ^ : i'eie\-anl fimi: l i o n a l e x p e r i e n c e w a s
s i g i ! d i ( ,ni[U' r e h i t e d ti* p j o h l e n i s i d n i t i l ' i e d i n t h e <irea u l i n t e r n a l n h i t i a g e -
n i i ' i i i - >p!'i i l i e a l l w > n h j e i : t s w i t h g i v . i i e r [ ) r o d u i ! i o n •'npi T a t i n n s o r l i n a n i e /
accciimlin;.; (!xp!!ririu,e w v j w signHii .nill\ nuiri' l i k e K t n ide!itif\' intei'iial
p i ' o h l e n i s i h i i i i w n r e o i l i e r s n l i j e c t s . I n l o t a l , \<\ i e l a t i ( a i s h i p s w e j e l e s l n d ,
w i t h o n l \ - t h e 2 r e p o i i i u i a h o \ e f o u n d t o h e si<_;nih( a n L
Di-^pile ihesi' p r r d o n i j n a i i t l \ ' null ri-sul!s, w e w a n i r d l o^iiard a.^ainst
I o n m i i t t i n , ^ a i \ ' p e II I'lTor, W e t h i ; r e l i n - e c a i r n l a l e d i h e s l i d i s l i c a ! pov\'i:r o f
i h e s t ! r r < _ ; r e ^ s i o n e q n a t i o n ^ l o r s [ n a l i - l o a n c d i i i r n e t l e ! t s i / r ' ~ (( j j h e j i . ! f i ; ^ H ) ,
s e t t i i i L : < i i p h a H q u a l t o , 1 0 . O n t h e h a s i , s <A\ WM-:-^-- i r . s i d K . w>> i v l a x e d f h e
pi'o[iahilit\ o ln I v p e Irrroi- [i.e,, falsel\' i-eji-diiii; t h en u l l h \ p o t ! i e s i s ] l o A l
l o r r d n i e t h e p r o l i a h i l i l v o f i . d s n l y a c r i ' p l i n g I h r n u l l h \ - p o l l i e s i s l o .n,'i. T h i s
a i l j i i s l n i i ' i i l r e s u l t e d ii! o n e U i o r e \ - a r i a h h ' ! ) e h i ; ; a s j i s n i h j . i i i t p r e d u J o r : fi~
r i a n f p . u i d a f i f j u n i in;_; e \ [ i e r i e n i e h a d a M ^ n H i c a n t , n ^ t ; , ! ! i \ e e f t c f J o i i s u h -
]('(1s" i d e n t i l \ - i n t ^ f i D a n r t ! a i u l .K i (.luntiii^ p r o h h i n i . s , •^lli^ t n r t h e r r e s u N w a s
isii o l l , s i ^ l e n ! w i l h M \ p o l h r s i s j h , a n d i ; i \ a M i t h a ! I h e {stliei' M ' ; ^ ^ n ' s s i o n r e s i d i s
o h t r i i n e d w e r e l o o M a l t r r e d t o r x e e i - d w l i a t m i ' . J i t Ise^ e . x | ) r ( h - d h \ c h a m e , w e
i o i i ( J u d e d d i a l o i n d a t a a L ^ o |.iile(i t o s u p p o r t l U [)ol|-ij'si> l h ,
h u r l h e r n i o n ! . r i i s i d l s i n t h e c a n o i i i c . i i a i i a l v s e s f o r H v ' p o t h e s i s i: o l l e r t H l
n o s u p p o r t l o r a ( o ! U i e ( , t i o n h e t w e e n helii^t s t i a n t u r e ^ a f u i d i e liiivi \\iy\\ o l
p e i H , e d i t i o n ' - r e t l e ( t ^ d i n e i t l i c ] - i d t o w i M l ictii i i s r d ( a ' p r o h l e i n ^ i d > ' u t i h e d . I n
h i s s t i i d v , V V a i s h ( 1 ! I 8 H : H;!,"I) l o n n d o n l \ s i a t i i T ' - d s i i p p o r l l o r t h e l e l . i i i o n -
s l i i p lii!tv\-eiiii b e l i e f s i n i c t u r e a n d p r o h l e i n s i d e n t i f i e d ,
i a h l e i ' j , i \ - e s r e s i d t s t o I' h ' s [ > , o t I ! \ ' p o t i i e s i ^ ' 1 , (.(MM e r u i u u t t u ' r i a r | - o w -
u e s s o rh r e a d d l o ! pei't ( ' [ j t i o n s w h e t t ( h el i r i i a d t h o t I n n ; t i o u i d e x p r r i e u i . e i s
M i n t r o l l e d . C ) h s e r \ a l i o n a l L ; o a l s d o a r l v h a d m a j o r iiftei:t,s o n h o w . s u i i j e c i s
a n a l w . i ' d t h e s e i tise.s, p a r t i i i i l a r l v i n l e r m s o | p r o b l e m i d e u t i l i c a l i o n . S i i b -
jej I s w h o w e r e a s k e d l o i d e i i t d v n i n h i j ) l e p r o h i e n i s a t t e n d ^ - d t o i n o i ' e iii((jr-
n i a t i i i n i n l a o n ; a n ' i i s a n d i d e n t i f i e d fiioi'e ( n ' o h h ' n ^ s i u u i o r r area--. T I K ^ S C

I ABI.K 1
Results i)f RRi^fKssion Analyses of Prmlictor Variables on Breadth
<)f Perceptions''
liifiunKilion ,\t!rnilf(l lo Priihlfnis Idi'iilitinl

l.)! A m c m i i l \ u n i b f i i^t A r e a s lul^il NtiirshiT \iinibt>i u

'*"/••- f)lil
728 Academy of Management Journal June

TABLE 2
Results of Regression Analyses of Direction of Perceptions on Areas of
Work Experience and Observational Goals"
Finance and Acconnting Human Resources
Information Problems Information Problems
Independent Variables Attended To Identified Attended To Identified
Finance and accounting experience -.03 -.27-' -.26^ -.20
Human resources experience -.07 -.22 -.17 -.06
Research and development experience .07 -.05 .06 -.06
Production and operations experience -.03 .08 -.25 -.18^
Information systems experience .03 -.30*** -.06 -.11
General management experience .15 -.05 -.01 -.02
Marketing and sales experience .01 -.12 -.34* -.27*
Observational goal .24** .32*** .17^ .50***
Adjusted R^ .04 .19*** .08* .26***
" For all models, df= 7, 115. Standardized regression coefficients are shown.
^ p < .10
* p < .05
* * p < .01
*** p < .001
results clearly support the hypothesis. Surprisingly, the breadth of func-
tional experience did not predict the breadth of perceptions, either for in-
formation attended to or for problems identified.
We also conservatively tested Hypotheses 4a, 4b, and 5, which predict
interactive effects, by examining part canonical correlations controlling for
the main effects of either functional experience or belief structures. None of
the hypothesized interactions were found. Neither observational goals nor
salience significantly interacted with either functional experience or belief
structures to affect the direction of perceptions as measured by information
attended to or problems identified (part adjusted r^^ ranged from .02 to .17).
The interactions of salience and belief structure were also not significant
(part adjusted r / = .08 and .02).
The above results are summarized, with supported hypotheses starred,
in Figure lc. Findings supported only Hypothesis 3.
Switching to an exploratory mode, we next assessed the relationships
between all areas of functional experience and all areas of cognitive process-
ing with another set of regression analyses. These results (Table 2) show a
surprising lack of correspondence between experience and cognitive pro-
cessing in the same functional area. Only in finance and accounting were
experience and problems identified related, and then only at a marginal level
of significance. Furthermore, the relationship is not positive, as theory pre-
dicts, but negative. Other significant relationships pair logically uruelated
types of information or problems with specific areas of functional experi-
ence. Also telling is that most (7 of 9) of these marginally significant rela-
tionships are negative.
Since only 9 of the 70 tests of significance involved produced significant
results with p < .10, one interpretation of these results is that they are chance
1997 Beyer et al. 729

TABLE 2 (continued)
External Management Marketing Internal Management
Information Problems Information Problems Information Problems
Attended To Identified Attended To Identified Attended To Identified
.09 .04 .23 -.05 -.01 .08
.05 .02 .09 -.09 .07 -.11
.07 .01 .25* .06 .01 -.03
-.09 .00 .04 .08 -.18 .18
.15 -.23* .07 -.09 .00 -.07
.05 -.10 .27* -.08 .04 .04
.04 .01 .15 -.06 -.12 -.08
.14 .51*** .17+ .66*** .14 .43**
.00 .21*** .04 .40*** .01 .18**

relationships. We again conducted a power analysis to decide whether we


had adequate statistical power in the regressions reported in Tahle 2 to
detect medium-sized effects. For prohlems identified, the power analysis
indicated that we had adequate power to detect relationships. For informa-
tion attended to, although the power analysis indicated that only with an
alpha value of .80 would we have sufficient power to detect the hypoth-
esized effects, we did ohtain several significant results.
If we accept these scattered results as nonrandom, they show that func-
tional experience tends to focus cognitive processes away from some unre-
lated areas, rather than direct attention to related areas. For example, suh-
jects with information systems experience identified fewer prohlems in hoth
finance and accounting and in external management than other suhjects.
Similarly, marketing and sales work experience did not significantly predict
attention to information in the marketing and sales area, hut it did have a
significant, negative impact on the information attended to in the area of
human resources. Production and operations experience was also negatively
related to the information attended to and prohlems identified in human
resources. More puzzling are the results for finance and accounting experi-
ence, which is negatively related at the .10 level hoth to the prohlems iden-
tified in finance and accounting and to the information attended to in human
resources. The only positive relationships are for the marketing information
attended to, which is positively related to research and development expe-
rience and general management experience. Again, there is no ohvious
match hetween experience and perceptions.
Tahle 2 also provides more specificity on the effects of ohservational
goals on the direction of perceptions in all functional areas. The results show
that ohservational goals hoth influenced the numher of prohlems identified
in all functional areas and influenced the information attended to for finance
and accounting, human resources, and marketing.
DISCUSSION
The purpose of this study was not only to see which of the results from
the Dearborn and Simon (1958) and the Walsh (1988) studies could he rep-
730 Academy of Management Journal June

Heated, but also to extend those studies by investigating more fully the links
between functional experience, belief structures, and selective perception.
To that end, we added observational goals and salience of belief structure as
additional predictors in our study to represent methodological factors that
could have affected subjects' information processing and problem identifi-
cation in the prior two studies. We also investigated possible interactions of
the main predictor variables and measured selective perception by both the
breadth and direction of subjects' information processing.
As might be expected in a study that assesses a comprehensive model of
all possible relationships, our results did not support all our hypotheses.
None of the interactive effects hypothesized in the model were found. In
particular, the salience of belief structures created by possible priming ef-
fects did not have significant interactive effects on selective perception. The
other aspect of experimental procedures investigated (namely, differences in
instructions to subjects in the two replicated studies) did, however, have
significant main effects. Our results showed clearly that observational goals
embodied in these instructions had direct effects on the breadth of percep-
tions—that is, on the amounts and type of information attended to and the
amounts and type of problems identified. Individuals who were encouraged
to think of multiple problems (as in the Walsh study) identified more prob-
lems and attended to more information than individuals whose assigned
goal was to identify a single problem (as in the Dearborn and Simon study).
Thus, we must conclude that the differences between the results of the two
earlier studies are at least partially attributable to methodological differ-
ences—specifically, to how many problems subjects were encouraged to
identify.
Further analyses—in which we decomposed work experience, informa-
tion attended to, and problems identified by functional area—revealed that
experience predicted only a few areas of information attended to or problems
identified. Although these results are scattered, they are interesting in that
logically related areas of work experience did not predict problems identi-
fied or information attended to. Rather, unrelated areas of experience tended
to predict both problems identified and information attended to, suggesting
that functional experience does not increase managers' attention to related
information but instead tends to restrict the areas of information to which
they pay attention. These suggestive findings support the general conclusion
that has been drawn from the Dearborn and Sim^on results, that managers'
functional experience tends to narrow their cognitive processing. Our find-
ings differ, however, from the usual interpretations of selective perception
that posit functional experience as directing attention to rather than away
from related areas of information. In particular, they are consistent with
Dearborn and Simon's speculations about selective attention leading to peo-
ple's ignoring some stimuli.
Much of the literature on perception seems to rest on the assumption
that the cognitive processes that draw attention to a stimulus and those that
draw attention away from a stimulus are the same. Discussions often treat
1997 Beyer et al, 731

perception as a process in which the blocking of the perception of some


stimuli automatically directs greater attention to others. Researchers use
terms like perceptual filters (Starbuck & Milliken, 1988) and blind spots
(Porter, 1980; Zajac & Bazerman, 1991) to argue that more attention is di-
rected to some stimuli than to others. Usually, it is argued that the stimuli to
which perceivers devote more attention are those that are consistent with
existing schemata or familiar in terms of the perceivers' past experience (Day
& Lord, 1992; Hambrick & Mason, 1984; Kiesler & SprouU, 1982), especially
when the stimuli are complex or ambiguous (Bruner, 1957). It seems to
follow from these arguments that perceivers will tend to ignore, forget, or
discount information that is discrepant with existing beliefs (Kiesler &
SprouU, 1982). The attending-to and not-attending-to processes are thus as-
sumed to be two sides of the same general process.
Although it is sparser, some literature has specifically addressed the
issue of why people may fail to perceive—the process we have called selec-
tive imperception. In perhaps the most comprehensive treatment, Kiesler
and SprouU (1982) summarized a variety of situational reasons why people
may not attend to certain information: (1) the information may not meet a
threshold or aspiration level for what they need or expect—it may be deemed
inadequate, (2) they may already have detected related information that they
expect to serve their purposes, (3) other, competing information may be more
salient, (4) the strength of the signal for the information may be relatively
weak, or (5) the category into which the information appears to fall may be
inconsistent with what they are seeking (Fiske & Newberg, 1990; Fiske &
Taylor, 1991). Bayster and Ford (1995) suggested that managers tend to ig-
nore information that does not fit the way they have classified a problem in
terms of functions within their firms and associated characteristics of deci-
sion making in those functions. Other possibilities are that managers think
they can rely on others' expertise in the areas to which they are not attending
(Melone, 1994), or their attending to certain categories of information in the
past has not been reinforced (Brunderson & Sutcliffe, 1995), or their account-
ability for the decision is low (Fiske, 1993).
All of these suggested reasons for selective imperception are largely
external to perceivers and thus do not directly address what is happening
within them. Most of these reasons are not relevant to the experimental
situation used in this experiment or the previous studies it replicated. Only
the second and fifth reasons given above seem to have plausibility in a
situation in which M.B.A. student subjects are working alone to solve an
unfamiliar case. Also, all of these suggested reasons for imperception apply
to the direction of perceptions, and not to the question of their general
breadth.
Another possibility, one that seems more applicable in an experimental
situation such as we used, is that some process that is internal to the person
doing the perceiving leads to selective iiTiperception. We found two theo-
retical explanations in the literature for such an internal process: the elimi-
nation of aspects (Tversky, 1972) and the switching of cognitive gears (Louis
732 Academy of Management Journal June

& Sutton, 1991; White & Carlston, 1983). In the first, the narrowing of indi-
viduals' perceptions operates more by their eliminating single aspects (Tver-
sky, 1972) they consider peripheral than by their focusing on matters they
consider central. Such an explanation is consistent with Walsh's contention
that managers are not such selective perceivers as had been thought. They
may keep a reasonable range of options open and eliminate options one at a
time rather than attend to only a few aspects that fit prior beliefs. This
explanation is consistent with most of the negative relationships found in
Table 2.
The third possibility is that people switch between actively attending to
schema-consistent and schema-inconsistent information (White & Carlston,
1983), When confronted with complex stimuli, they tend to look, in what is
called controlled information processing, for schema-consistent information
first. Once they have confirmed their schemata, they turn their attention
away from schema-consistent information to schema-inconsistent informa-
tion, but continue to monitor schema-consistent information in an auto-
matic, less demanding process. It therefore seems possible that managers
may pay less attention to information consistent with schemata derived from
their functional experience after an initial period in which they have at-
tended to enough of that information to confirm their functional schemata.
It also seems possible that they may need relatively little information to
confirm their schemata. Such a process would explain our results from fi-
nance and accounting, in which the subjects with the most experience in
these fields paid less attention to information in these areas. These argu-
ments, however, also seem to address the direction of perceptions more than
they address their breadth.
Although the explanations offered above may also pertain to the breadth
of perceptions to a degree, they are undoubtedly not the whole story. It
seems likely that the internal dispositions or traits of a perceiver also affect
perceptions. One such trait that bears examination is cognitive complexity.
In an experimental mode, we included cognitive complexity in this study
and found it had strong effects on the breadth of perceptions.^ Thus, an
especially promising avenue for addressing the breadth of perceptions in
future research would be to further explore the effects of cognitive complex-
ity on selective perception or imperception and to work at identifying other
individual dispositions that seem likely to affect how broadly or narrowly
people perceive.
Walsh's (1988) study and this study considered three aspects of selec-
tive perception: the direction of information sought, the information at-
tended to, and the problems identified. We did not find any consistent

'' Cognitive complexity was also positively related (p < .10) with information attended to in
four areas, and with problems identified in two areas. Inclusion of cognitive complexity in-
creased most of the adjusted R^s predicting tha direction of perceptions but did not change the
general pattern of results for the effects of work experience and observational goals. These
results are available from the first author.
1997 Beyer et al. 733

effects of belief structures on any of these measures. There are at least two
possible reasons why: First, there may actually be no such effects. Walsh also
failed to find consistent relationships between belief structures and the di-
rection of perceptions. Second, it may be extremely difficult—perhaps im-
possible—to measure belief structures adequately (Rosman, Lubatkin, &
O'Neill, 1994). Although it is appealing to posit a belief structure that chan-
nels problem identification and the cognitive processing of information,
neither we nor Walsh (1988) were able to demonstrate such a relationship
using two different sets of carefully developed measures.
In addition, this study yielded strong direct effects of observational
goals on the breadth of perceptions. At least in experiments, people's per-
ceptions of problems and information are responsive to instructions.
Like the studies replicated, this study derives both strengths and limi-
tations from its research method, a controlled but contrived experimental
situation. A strength of the present study that the earlier studies lacked is
that we employed experimental controls that enabled us to test the theory
underlying the selective perception construct more comprehensively and
rigorously than the past studies had done. This type of research cannot,
however, determine whether practicing managers behave like our research
subjects when actually making crucial decisions.
In addition, our subject pool differed somewhat from that of the repli-
cated studies in years of managerial experience. We cannot be sure how
many years of experience their subjects had. Dearborn and Simon only de-
scribed their subjects as "executives," and Walsh's subjects averaged 38
years of age while ours averaged 28 years. Despite these differences, how-
ever, our results are actually quite similar to those of the earlier studies. Like
Dearborn and Simon, we found functional experience was sometimes re-
lated to the types of problems subjects identified. However, in our results
nonsignificant relationships were so much more numerous than significant
ones that we were forced to conclude that our results were equivocal at best
and did not give clear evidence for these relationships. Also, like Walsh, we
failed to find a link between belief structures and functional experience and
found only scattered relationships between belief structures and information
processing. Furthermore, his and our subjects identified similar numbers of
problems (means were 7.12 and 7.47, respectively) and attended to similar
numbers of pieces of information (means were 25.09 and 20.76, respec-
tively)."^ Thus, it seems unlikely that differences in status and age between
our subjects and his produced any differences in effects between the two
studies. The main difference between our results and his are the strong and

" In addition, we explicitly tested the possibility that an M.B.A. education affected subjects'
responses by looking for differences in our results between first- and second-year students in
our sample. The first-year students participated in this experiment before their classes began,
during orientation week. The second group participated in the last semester of what is usually
a two-year program. There were no significant differences in results as a whole between the two
classes of students.
734 Academy of Management Journal June

consistent effects of a methodological factor (observational goal) we con-


trolled. Observational goals not only help to explain the differences in the
results of the two earlier studies but, as other research has suggested, seem
likely to influence the information processing of managers regardless of
experience. A second difference is that our study uncovered evidence of
selective imperception, which could not be detected by the statistical pro-
cedures used in {he prior studies. Thus, we do not know whether or not
imperception was present in the managerial samples of respondents they
used.

CONCLUSIONS

The results of Dearhorn and Simon's (1958) study of selective percep-


tion produced a widespread helief among management scholars that man-
agers are selective information processors. Our results suggest that managers'
information processing is somewhat influenced hy their functional experi-
ence, hut not necessarily in the ways previously theorized and investigated.
Furthermore, the cognitive path through which such influence might occur
remains to he demonstrated. M.B.A. students with managerial work experi-
ence exercised selective perception when constrained by the ohservational
goals given them, but they were more catholic information processors when
their observational goals called for a broader perspective. When their func-
tional experience did influence their information processing, it was more
likely to direct that attention away from unrelated areas of information than
toward related areas, suggesting that managers may narrow their perceptions
through selective imperception.

REFERENCES
Aronson, E., Ellsworth, P. C, Carlsmith, J. M., & Gonzales, M. H. 1990. Methods of research in
social psychology. New York: McGraw-Hill.
Bayster, P. G., & Ford, C. M. 1997. The impact of functional classification schema on managerial
decision processes. Journal of Managerial Issues, 9: In press.
Brunderson, J. S., & Sutcliffe, K. M. 1995. Rethinking the relationship hetween work history
and selective perception. Paper presented at the annual meeting of the Academy of Man-
agement, Vancouver.
Bruner, J. S. 1957. On perceptual readiness. Psychological Review, 64: 123-152.
Gohen, J. C. 1988. Statistical power analysis for the behavioral sciences. Hillsdale, NJ:
Erlbaum.
Day, D. V., & Lord, R. G. 1992. Expertise and problem categorization: The role of expert pro-
cessing in organizational sense-making. Journal of Management Studies, 29: 35-47.
Dearborn, D. C., & Simon, H. A. 1958. Selective perception: A note on the departmental iden-
tifications of executives. Sociometry, 21: 140-144.
De Nisi, A. S., Cafferty, T. P., & Meglino, B. M. 1984. A cognitive view of the performance
appraisal process: A model and research propositions. Organizational Behavior and Hu-
man Performance, 33: 360-396.
1997 Beyer et al. 735

Devine, P. G., Monteith, M. J., Zuwerink, J. R., & Elliott, A. J. 1991. Prejudice with and without
compunction. Journal of Personality and Social Psychology, 60: 817-830.
Dweck, C. S., & Leggett, E. L. 1988. A social-cognitive approach to motivation and personality.
Psychological Review, 95: 256-273.
Fiske, S. T. 1993. Social cognition and social perception. In M. R. Rosenzweig & L. W. Porter
(Eds.), Annual review of psychology, vol. 44: 155-194. Palo Alto, CA: Annual Reviews.
Fiske, S., & Neuberg, S. 1990. A continuum of impression formation, from category-based to
individuating processes: Influences of information and motivation on attention and inter-
pretation. In M. Zanna (Ed.), Advances in experimental social psychology, vol. 23: 1-74.
New York: Academic Press.
Fiske, S., & Taylor, S. E. 1991. Social cognition. New York: McGraw-Hill.
Glick, W. H., Miller, G. G., & Huber, G. P. 1993. The impact of upper-echelon diversity on
organizational performance. In G. P. Huber & W. H. Glick (Eds.), Organizational change
and redesign: Ideas and insights for improving performance: 176-214. New York: Ox-
ford University Press.
Hambrick, D. G., & Mason, P. A. 1984. Upper echelons: The organization as a reflection of its top
managers. Academy of Management Review, 9: 193-206.
Hoch, S. J. 1984. Availability and inference in predictive judgment. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 10: 649-662.
Ilgen, D. R., & Feldman, J. M. 1983. Performance appraisal: A process approach. In L. L. Gum-
mings & B. M. Staw (Eds.), Research in organizational behavior, vol. 5: 141-197. Green-
wich, GT: JAI Press.
Kiesler, S., & Sproull, L. 1982. Managerial response to changing environments: Perspectives and
problem sensing from social cognition. Administrative Science Quarterly, 37: 548-570.
Latham, G. P., Erez, M., & Locke, E. A. 1988. Resolving scientific disputes by the joint design of
crucial experiments by the antagonists: Application to the Erez-Latham dispute regarding
participation in goal setting. Journal of Applied Psychology, 73: 753-772.
Lord, R. G. 1985. An information-processing approach to social perceptions, leadership and
behavioral measurement in organizations. In L. L. Gummings & B. M. Staw (Eds.), Research
in organizational behavior, vol. 7: 87-128. Greenwich, GT: JAI Press.
Louis, M. R., & Sutton, R. I. 1991. Switching cognitive gears: From habits of mind to action
thinking. Human Relations, 44: 55-61.
Melone, N. P. 1994. Reasoning in the executive suite: The influence of role/experience-based
expertise on decision processes of corporate executives. Organization Science, 5: 438—455.
Mintzberg, H. A., Raisinghani, D., & Th6oret, A. 1976. The structure of unstructured decision
processes. Administrative Science Quarterly, 21: 246-275.
Mulaik, S. A. 1972. The foundations of factor analysis. New York: McGraw-Hill.
Neuberg, S. L. 1989. The goal of forming accurate impressions during social interactions: At-
tenuating the impact of negative expectancies. Journal of Personality and Social Psychol-
ogy, 56: 374-86.
Pavelchek, M. A. 1989. Piecemeal and category-based evaluation: An idiographic analysis. Jour-
nal of Personality and Social Psychology, 56: 354-363.
Porter, M. 1980. Competitive strategy: Techniques for analyzing industries and competitors.
New York: Free Press.
Prahalad, G. K., & Bettis, R. A. 1986. The dominant logic: A new linkage between diversity and
performance. Strategic Management Journal, 7: 485-501.
736 Academy of Management Journal June

Quinn, R,, & Rohrbaugh, J. 1983. A spatial model of effectiveness criteria: Towards a competing
values approach to organizational analysis. Management Science, 29: 363-377.
Robinson, E., & Pearce, J. 1988. Planned patterns of strategic behavior and their relationship to
business unit performance. Strategic Management Journal, 9: 43-60.
Rosman, A., Lubatkin, M., & O'Neill, H. 1994. Rigidity in decision behaviors: A within-subject
test of information acquisition using strategic and financial informational cues. Academy
of Management Journal, 37: 1017-1033.
Shrout, P. E., & Fleiss, J. L. 1979. Interclass correlations: Uses in assessing rater reliability.
Psychological Bulletin, 86: 420-428.
Squire, L. R., Knowlton, B., & Musen, J. 1993. The structure and organization of memory. In
M. R. Rosenzweig & L. W. Porter (Eds.), Annual review of psychology, vol. 44: 453-495.
Palo Alto, CA: Annual Reviews.
Starbuck, W. H., & Milliken, F. J. 1988. Executives' perceptual filters: What they notice and how
they make sense. In D. C. Hambrick (Ed.), The executive effect: Concepts and methods for
studying top managers: 35-65. Greenwich, CT: JAI Press.
Tversky, A. 1972. Elimination by aspects: A theory of choice. Psychological Review, 79: 281-
299.
Tversky, A., & Kahneman, D. 1974. Judgment under uncertainty: Heuristics and biases. Science,
185: 1124-1131.
Walsh, J. P. 1988. Selectivity and selective perception: An investigation of managers' belief
structures and information processing. Academy of Management Journal, 31: 873-896.
Weick, K. E. 1979. Cognitive processes in organizations. In B. Staw (Ed.), Research in organ-
izational behavior, vol. 1: 41-74. Creenwich, CT: JAI Press.
White, A. D., & Carlston, D. E. 1983. Consequences of schemata for attention, impressions, and
recall in complex social interactions. Journal of Personality and Social Psychology, 45:
538-549.
Zajac, E. J., & Bazerman, M. H. 1991. Blind spots in industry and competitor analysis: Implica-
tions of inter-firm (mis)perceptions for strategic decisions. Academy of Management Be-
view, 16: 37-36.

Janice M. Beyer is the Harkins & Company Centennial Chair in Management, a profes-
sor of sociology, and the Director of the Center for Organizations Research at the Uni-
versity of Texas in Austin. She earned her Ph.D. degree from the New York State School
of Industrial and Labor Relations at Cornell University. Her current research interests
center on socialization, total quality management, and cultures in work organizations.
Prithviraj Chattopadhyay is an assistant professor in the School of Public Affairs and
Administration at Western Michigan University. He received his Ph.D. degree in man-
agement from the University of Texas at Austin. His research interests include demog-
raphy, citizenship behavior, and managerial cognition.
Elizabeth George is an assistant professor in the Haworth College of Business at West-
ern Michigan University. She received her Ph.D. degree in management from the Uni-
versity of Texas at Austin. Her research interests include employment externalization,
institutionalization of employment practices, and managerial cognition.
William H. Glick is the chair of the Department of Management and a Dean's Council
of 100 Distinguished Scholar at Arizona State University. He received his Ph.D. degree
from the University of California, Berkeley. His main teaching and research interests are
in the areas of job and organizational design, business process redesign, and diversity
in top management teams.
dt ogilvie is an assistant professor of strategic management on tbe Faculty of tbe Cradu-
1997 Beyer et al. 7

ate School of Management at Rutgers University. She earned her Ph.D. degree in stra-
tegic management from the University of Texas at Austin. Her primary research is in the
area of strategic decision making and the use of creativity to enhance business decision
making. Her research interests also include executive leadership strategies, assessing
environmental dimensions, and technology transfer and countertrade issues in the
People's Republic of China.
Dulce Pugliese is a Ph.D. candidate in management at the Graduate School of Business
of the University of Texas at Austin. She received her M.D. and M.B.A. degrees from the
Universidade Federal do Rio de Janeiro. Her research interests include the development
of new organizational forms and new organizational populations, and the health care
industry.

You might also like