environment and sdmprocess review 2014

8/10/2019 Environment and SDMProcess REVIEW 2014

1/25

The Influence of Context on the StrategicDecision-Making Process: A Review of

the Literature

Neil Gareth Shepherd and John Maynard RuddMarketing Group, Aston Business School, Aston University, Aston Triangle, Birmingham B4 7ET, UK

Corresponding author email:[email protected]

This paper critically reviews the strategic decision-making process literature, with aspecific focus on the effects of context. Context refers to the top management team,

strategic decision-specific characteristics, the external environment and firm charac-

teristics. This literature review also develops an illustrative framework that incorpo-rates these four different categories of contextual variables that influence the strategicdecision-making process. As a result of the variety and pervasiveness of contextualvariables featured within the literature, a comprehensive and up-to-date review is

essential for organizing and synthesizing the extant literature to explicate an agenda forfuture research. The purpose of this literature review is threefold: first, to criticallyreview the strategic decision-making process literature to highlight the underlying

themes, issues, tensions and debates in the field; second, to identify the opportunitiesfor future theory development; and third, to state the methodological implications

arising from this review.

Introduction

Strategic decisions (SDs) can be ill-structured, non-

routine, uncertain and pervasive. They cut across

organizational functions, entail a significant financial

outlay, and have profound, long-term implications

for the organization (Eisenhardt and Zbaracki 1992;

Mintzberg et al. 1976; Shrivastava and Grant 1985).

While SDs are not always entirely different from

other organizational decisions, they are towards one

end of a continuum, at the other end of which are the

trivial everyday questions (Hickson et al. 1986, p.

27). The strategic decision-making process (SDMP)

is described as a set of different characteristics, such

as rational, comprehensive, political; or as a

sequence of activities (Goll and Rasheed 2005) that

involves information gathering, developing alterna-tives and choosing among alternatives (Wally and

Baum 1994). Strategic decision-making process

research is of great importance, because the insights

that it provides can improve the effectiveness of SDs

made by executives, which ultimately contribute to

the success of organizations.

Context refers to the top management team

(TMT), strategic decision-specific characteristics,

the external environment and firm characteristics

(Sharfman and Dean 1997b), and the lack of a sys-

tematic treatment of contextual variables has resulted

in an incomplete, and perhaps inaccurate, picture ofSDM (strategic decision-making) (Hough and

White 2003, p. 488). Most existing SDMP studies

have adopted an incremental approach to theory

development, and focused only on a limited number

of contextual variables. This has resulted in a frag-

mented understanding, and left SDMP scholars

unable to identify the key contextual influences on

The authors would like to thank Kamel Mellahi, Editor-in-Chief, and the anonymous reviewers for their helpful andinsightful comments on earlier drafts of this work.

bs_bs_banner

International Journal of Management Reviews, Vol. 16, 340364 (2014)

DOI: 10.1111/ijmr.12023

2013 British Academy of Management and John Wiley & Sons Ltd. Published by John Wiley & Sons Ltd, 9600 GarsingtonRoad, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA
mailto:[email protected]:[email protected]


2/25

the SDMP (Papadakis and Barwise 1997). Inconsist-

encies among existing studies highlight the need for

future research to pay closer attention to context

(Elbanna and Child 2007b). For example, the effects

of SDMP comprehensiveness on organizational

performance remain unclear, with Fredrickson and

Mitchell (1984) finding negative effects, andBourgeois and Eisenhardt (1988) finding positive

effects. Closer scrutiny of context will reveal insights

that can help to reconcile such contradictory results.

Elbanna (2006, p. 14) emphasizes the need for this

review, stating that we still know little about the role

of other contextual variables in the SDMP, and that

a next logical step in this line of critical review

would be to review the role of contextual variables in

the SDMP. Moreover, a priority for SDMP research

is to identify the extent to which variance in the

characteristics of the SDMP is explained by the

context in which it takes place (Papadakis et al.1998; Rajagopalan et al. 1997). Similarly, Elbanna

and Child (2007b) and Pettigrew (2003) emphasize

that the rationality of the SDMP cannot be fully

understood without comprehension of its context.

Furthermore, it is vitally important to examine inter-

actions between contextual variables and SDMP

characteristics, because such interactions have sig-

nificant implications for SDMP outcomes, such as

the overall quality of a decision (Papadakis et al.

2010).

This paper critically reviews the SDMP literature

with a specific focus on the effects of contextualvariables, to provide an in-depth analysis of the

underlying themes, issues, tensions and debates in

the domain, and to identify priorities for future

research, together with the important methodological

implications. An illustrative framework of the con-

textual variables and their relationships with SDMP

characteristics and outcomes is presented (Figure 1).

The illustrative framework provides a scheme around

which this review is structured. Such a review

scheme allows the systematic identification of

themes and contributions, and allows the similarities

and discrepancies from such a diverse set of studies

to be discerned (Ginsberg and Venkatraman 1985).

Approach to the review

To ensure inclusiveness with regard to the literature

reviewed, this paper adopts an ontological framework

whereby the concept of a decision is viewed as valu-

able for understanding aspects of organizational

behaviour (Miller 2010). However, certain authors

(e.g. Mintzberg and Waters 1990) suggest that exam-

ining decisions can be a hindrance to understanding

organizational processes, because individual deci-

sions can be troublesome to identify. Actions can

occur without a formal decision having been made,

and organizations can take a particular course ofaction in response to the external environment, rather

than as a result of a systematic decision process.

Ultimately, whether decisions are suitable subjects

for empirical enquiry or not, and whether researchers

can obtain objective knowledge of decision pro-

cesses, rests upon the ontology of the researcher

(Pettigrew 1990). The position taken in this paper is

that an understanding of SDMPs is possible, and is

useful for explaining differences in organizational

performance. However, such research is challenging,

in great part because of the complex influence of

context.To conduct this review, and in keeping with

prior classifications (e.g. Papadakis et al. 1998;

Rajagopalan et al. 1993; 1997), we derived four cat-

egories of contextual variables from an extensive lit-

erature search: the TMT; SD-specific characteristics;

the external environment; and firm characteristics.

Each has either a direct effect on the characteristics

of the SDMP or a moderating effect on the relation-

ship between SDMP characteristics and SDMP out-

comes (see Figure 1). Some studies also specify

the direct effects of contextual variables on SDMP

outcomes.Variables pertaining to each of the four categories

were identified through keyword searches of top

peer-reviewed academic journals in the ProQuest,

EBSCO, Emerald Full Text, JSTOR Business,

Science Direct and PsycArticles databases. Each

article was then allocated to one of the four catego-

ries of contextual variables. To ensure reliability, the

classification decisions were independently verified

by a senior academic familiar with the subject. To

assure the completeness of the search, the Social

Science Citation Index Journal Impact Factor was

used to identify top peer-reviewed journals that

commonly address the topic of strategic decision-

making; these journals were then searched indivi-

dually for articles featuring contextual variables.

Manual searches of multiple reference lists were also

conducted, and an electronic library catalogue was

searched to identify relevant books.

The search procedures generated a substantial

working list of articles. This list was narrowed to

specifically relevant articles by applying three strict

Context in Strategic Decision-Making 341

2013 British Academy of Management and John Wiley & Sons Ltd.


3/25

SDM

PCharacteristics

SDMPOutcomes

ContextualVariab

les

Rationality

Compreh

ensiveness

Political

behaviour/politicization

Intuition/intuitivesynthesis

Lateralcommunication

Hierarchicaldecentralization

Problem

solvingdissension

Useoffinancialreporting

Ruleform

alization

Flexibility

Agreeme

ntaboutcomprehensiveness

Changei

ncomprehensiveness

Conflict

Participation

Levelof

aggression

Toleranceofambiguity/uncertainty

Assessment/perceptionsofrisk

Technocratic

Sporadic

Fluid

Constricted

Effectiveness

Quality

Speed

Commitment

Erraticstrategic

decisions

Organizational

Performance

Decisiveness

Newproduct

performance/quality

TopManagementTea

m

(Table1)

Demographics(tenure,

education,diversity,age)

Cognitivediversity/style

Personality

Other(meta-cognition,potency,

polychronicity,aggressiv

e

philosophy)

StrategicDecisionSpec

ific

Characteristics

(Table2)

Matter

Uncertainty

Motive

Importance

Timepressure

TheExternalEnvironm

ent

(Table3)

Dynamism

Velocity

Munificence

Hostility

Instability

Uncertainty

FirmCharacteristics

(Table4)

Powercentralization

Structure

Size

Performance/Slackresources

Other(externalcontrol,

corporatecontrol,formal

planningsystems)

Directeffects

Moderatingeffects

Figure

1.Contextua

lvariab

lesinSD

MPresearch:anillustrative

framewor

kofe

xtantresearch

342 N.G. Shepherd and J.M. Rudd



4/25

criteria, and articles were included in the final list

only if they: (i) explicitly examine the direct effects

of a particular contextual variable on SDMP charac-

teristics, or (ii) explicitly examine the effects of a

particular contextual variable on the relationship

between SDMP characteristics and SDMP outcomes,

or (iii) explicitly examine the effects of a particularcontextual variable on SDMP decision-level out-

comes (e.g. decision speed, decision quality).

Articles were excluded if the main focus was not the

process of strategic decision-making.

This review is structured as follows. First, we criti-

cally review the empirical literature relating to each

of the four categories of contextual variables to high-

light the underlying themes, issues, tensions and

debates in the field. We structure this review accord-

ing to studies examining the direct effects of contex-

tual variables on SDMP characteristics, to those

examining the direct effects on SDMP outcomes andto those examining the moderating effects of contex-

tual variables on the relationship between SDMP

characteristics and outcomes. Second, we discuss the

priorities for future theory development, and, third,

we state the methodological implications arising

from this review.

The top management team

The TMT is the dominant coalition of the most senior

executives who have responsibility for setting theoverall direction of the organization (Hambrick and

Mason 1984). Articles reviewed here are predicated

on the concept of bounded rationality, whereby SDs

are the product of behavioural influences, rather than

economic utility maximizing processes (Cyert and

March 1963; March and Simon 1958). Bounded

rationality acknowledges inherent cognitive limita-

tions of decision-makers that restrict their ability to

collect and analyse all relevant information and iden-

tify all possible alternatives (Griffith et al. 2012).

The TMT variables reviewed in this section include

demographic (tenure, education and diversity) and

psychometric variables (cognitive diversity, cognitive

style and personality) (see Figure 1 and Table 1). As

well, TMT variables feature predominantly as ante-

cedents of SDMP characteristics and outcomes, with

no studies examining their moderating influence on

the SDMP characteristicsoutcomes relationships.

In the current section, we review articles that

examine the direct effects of TMT demographic and

psychometric variables on SDMP characteristics, as

well as articles that examine the direct effects of

these TMT variables on SDMP outcomes. Consistent

with prior treatment in the literature (e.g. Papadakis

and Barwise 2002; Papadakis et al. 1998), CEO

variables are included in this section, because the

CEO is cited as the most influential TMT member

(Hambrick and Mason 1984).

Direct effects of TMT demographic variables on

SDMP characteristics

Tenure. Tenure of the TMT and the CEO,

operationalized as length of service with an organi-

zation, has been found to influence the level of

rationality, comprehensiveness and the extent of

middle management participation in the SDMP. Goll

and Rasheed (2005) found that long-tenured TMTs

adopt rational SDMPs they engage in continuous

proactive searches, undertake extensive analysis andconduct formal planning. Empirical evidence also

indicates that, as the tenure of the TMT increases,

so too does comprehensiveness. Fredrickson and

Iaquinto (1989) term this phenomenon creeping

rationality, and, as tenure increases,TMTs engage in

ever more thorough situation diagnosis, generation

and evaluation of alternatives, and integration of the

decision into the overall strategy of the firm.

Though middle management in organizations led

by long-tenured CEOs have a higher level of involve-

ment in the SDMP (termed hierarchical decentrali-

zation), CEO tenure does not significantly affectother important and frequently studied SDMP char-

acteristics such as comprehensiveness (Papadakis

and Barwise 2002; Papadakis et al. 1998). Hence,

the CEO may be the most powerful member of the

TMT (Hambrick and Mason 1984), yet the effect of

CEO tenure on SDMP comprehensiveness appears to

be less influential than that of TMTs.

TMT education level. TMT education level (e.g.

high school, undergraduate degree, Masters degree)

influences rationality and comprehensiveness in

SDMP. Highly educated TMTs are more rational in

SDM, owing to the strengthened analytical ability

that results from increased education level (Goll and

Rasheed 2005). Similarly, Papadakis and Barwise

(2002) found that highly educated TMTs are more

comprehensive in situation diagnosis, alternatives

generation and evaluation, and integrating the

decision into the firms overall strategy. While TMT

education level influences rationality and com-

prehensiveness, Papadakis et al. (1998) determined




5/25


6/25

Forbes(2005)

Founder-managerage

andpriorventure

experience

98founder-managers

(entrepreneurs)ofInternet

start-ups

Field-based;cross-sectional;

on-lineandmailsurvey;

archivaldata;recentSDsmade

bythefirms

Regression

analysis

FirmsmakefasterSDswhenmanage

dbyolder

founder-managersandbythosewithpriorventure

experience

GollandRasheed

(2005)

TMTage,tenure

,and

educationleve

l

159Manufacturingfirms

Field-based;cross-sectional;mail

survey;archivaldata;actual

SDMPs

Regression

analysis

AveragetenureandeducationleveloftheTMTispositively

relatedtorationality,butaveragea

geisnot

Houghandogilvie

(2005)

Cognitivestyle

749executives

Laboratory-basedbehavioural

simulation

Structurale

quation

modelling

Executiveswhousebothintuitionandobjectiveinformation

makehigherqualitydecisions,whilethosewhotaketimeto

makesociallyacceptabledecisions

areindecisiveand

perceivedtobeineffective

Olsonetal.(2007a)

TMTcognitived

iversity

TMTsfrom85hospitals


survey;recentSDmadebythe

firms

Regression

analysis

Cognitivediversityhasapositiverelationshipwithtask

conflict,andcompetence-basedtru

ststrengthensthis

relationship.Taskconflictmediatestherelationshipbetween

cognitivediversityandSDunderstanding,commitmentand

quality

Olsonetal.(2007b)

TMTcognitived

iversity

252Chineseexecutives

Field-based;crosssectional;


firms

Regression

analysis

CognitivediversityhasanegativerelationshipwithSD

commitmentandquality,andtherelationshipsare

moderatedbyaffect-basedandcog

nition-basedtrust

Souitarisand

Maestro(2010)

TMTpolychronicity

129TMTsfromnewtechnology

ventures



firms

Regression

analysis

TMTpolychronicityhasapositiveef

fectonSDspeedanda

negativeeffectonSDMPcomprehensiveness.SDspeedand

comprehensivenesspartiallymedia

tetherelationship

betweenTMTpolychronicityandfinancialperformance

Mitchelletal.

(2011)

CEOmeta-cognition

64CEOsoftechnologyfirms

Field-basedexperiment;

hypotheticalchoices;mail

survey

Conjointan

alysis

TheSDsmadebyCEOswithgreatermetacognitiveexperience

arelesserratic

ClarkandMaggitti

(2012)

TMTpotency

54TMTsofhigh-technology

companies

Field-based;cross-sectional;

interviews;mailsurveyof

actualSDMPs

Regression

analysis

TMTpotencyispositivelyrelatedto

SDspeed.TMTpotency

partiallymediatestherelationshipbetweenTMTexperience,

knowledge,interactionprocessand

SDspeed




7/25

that CEO education level affects only the use of

financial reporting information. Highly educated

CEOs make greater use of financial reporting infor-

mation in the SDMP; however, CEO education level

does not significantly influence any other character-

istics of the SDMP, such as comprehensiveness

(Papadakis and Barwise 2002; Papadakis et al.1998).

TMT demographic diversity. TMT demographic

diversity refers to heterogeneity in the age, tenure,

experience and education in the TMT, and has only

weak direct effects on comprehensiveness defined

as thoroughness and inclusiveness in making and

integrating SDs (Simons et al. 1999). The same

study identified that interactions between job-related

demographic diversity measures (e.g. company

tenure diversity) and TMT debate (discussions

concerning how to approach the decision) are moresignificant predictors of comprehensiveness than

interactions between non-job-related diversity meas-

ures (e.g. age diversity) and TMT debate.

The findings of the studies featuring TMT and

CEO tenure, education and demographic diversity

reviewed here show that demographic variables such

as tenure and education do significantly affect char-

acteristics of the SDMP such as rationality, compre-

hensiveness, hierarchical decentralization and the

use of financial reporting information. However, job-

related TMT demographic diversity variables appear

to influence comprehensiveness more than non-job-related demographic diversity variables. It is also

apparent that TMT and CEO demographics exert dif-

ferent influences on SDMP characteristics. Certainly,

TMT demographic variables appear to be a stronger

influence on SDMP comprehensiveness than those of

the CEO.

A lack of significant results from studies examin-

ing CEO demographics brings into question the

focus on individual CEOs for explaining SDMP

characteristics. Hambrick and Mason (1984)

acknowledge that the TMT as a unit of analysis has

more explanatory power than the individual CEO,

because tasks such as SDM are rarely undertaken

exclusively by the CEO alone, but more often are

shared between members of the TMT. Also, the con-

struct validity of demographic variables has been

questioned in the literature, prompting calls to use

direct psychometric measures of TMT cognitive and

behavioural traits (e.g. cognitive diversity instead of

demographic diversity) to enhance the validity and

explanatory capability of research (e.g. Hough and

ogilvie 2005; Priemet al. 1999). While methodologi-

cally convenient, using demographic variables leaves

a black box (Lawrence 1997) of unexplained TMT

cognitions, values and perceptions that influence

SDs.

Direct effects of TMT psychometric variables on


Cognitive diversity. Cognitive diversity refers to

differences in the preferences and beliefs of

TMT members concerning the strategic goals and

priorities of the organization (Miller et al. 1998).

Cognitive diversity has been found to reduce com-

prehensiveness (Milleret al. 1998) and to increase

task conflict (Olsonet al. 2007a). Milleret al. (1998)

found that cognitively diverse TMTs are less com-

prehensive; instead, their SDMPs are characterized

by limited brainstorming, the consideration of anarrow range of alternatives and restricted use of

quantitative analyses. Cognitive diversity also causes

task conflict, and, when TMT members disagree

about the strategic goals and priorities of the organi-

zation, the SDMP is characterized by discord and

differences in judgment (Olson et al. 2007a). The

same study also found that, when TMT members

trust one anothers competence, the effects of cogni-

tive diversity on task conflict are stronger. Hence,

these two studies show that cognitive diversity

diminishes comprehensive analysis and the consid-

eration of multiple decision options and, instead,gives rise to conflict and disagreement during the

SDMP.

Cognitive style. Cognitive style describes how

people perceive, think, solve problems, learn, and

relate to each other (Hough and ogilvie 2005, p.

421). Using laboratory simulations and measuring

cognitive style by applying the MyersBriggs Type

Indicator, Henderson and Nutt (1980) and Nutt

(1993) discovered that executive cognitive style

influenced the level of aggression, tolerance of ambi-

guity and uncertainty, and assessment of risk in the

SDMP. Nutt (1990) also used a laboratory-based

simulation, measuring the cognitive style of 79

senior executives and 89 middle managers through

the MyersBriggs Type Indicator. The study shows

that cognitive style determines the perceived risk in

making the SD, as well as the likelihood of taking

strategic action. The study also indicates that cogni-

tive style determines the type of data and the mode of

data processing used in the SDMP. However, while




8/25

these studies of cognitive style highlight its potential

importance, it remains unclear how cognitive style

influences the SDMP outside laboratory conditions

in situations where multiple members of the TMT are

involved in the SDMP, rather than just one indi-

vidual. It is important for future research to deter-

mine whether the cognitive styles of TMT membersinteract to reinforce a particular style, or interact

such that a balance of styles is achieved (Hough and

ogilvie 2005). Also, the effects of cognitive style on

important and frequently studied SDMP characteris-

tics, such as comprehensiveness, are unknown.

Personality. Research has examined the effects of

CEO personality on the characteristics of the SDMP,

and has produced equivocal results. Miller et al.

(1988) found that CEOs with a higher need for

achievement (desire to attain success and accomplish

difficult tasks) are more rational in the SDMP(increased planning, systematic scanning of the envi-

ronment and having explicit strategies). However,

Papadakis et al. (1998) found no significant effects

of CEO need for achievement on any of the SDMP

characteristics. Papadakis et al. (1998) do find,

however, that CEO risk propensity (willingness to

take risks) diminishes rule formalization in the

SDMP (standardized procedures and processes), but

does not significantly affect any other SDMP char-

acteristic. Furthermore, Papadakis and Barwise

(2002) were unable to find any significant effects of

CEO need for achievement or risk propensity on anyof the characteristics of the SDMP. The contradictory

conclusions of Milleret al. (1988) and Papadakis

et al. (1998) concerning CEO need for achievement

could be attributable to differences in the size of the

organizations sampled. Milleret al. (1988) focused

on small organizations, where the influence of the

CEO is likely to be pervasive, and the average

number of employees in organizations in this sample

was 111, compared with an average number of

employees of 730 in the organizations in the

Papadakis et al. (1998) and Papadakis and Barwise

(2002) studies.

Other TMT variables. Other TMT variables fea-

tured in the literature include polychronicity and

aggressive philosophy. Souitaris and Maestro (2010)

studied TMT polychronicity, which refers to the

TMTs tendency to move focus simultaneously and

intermittently from one task to another. The results

show that TMT polychronicity leads to faster deci-

sions, but the TMTs are less comprehensive in their

SDMP. Also, an aggressive TMT philosophy (deter-

mination to beat competition) has been found to

result in increased SDMP comprehensiveness,

hierarchical decentralization, rule formalization and

lateral communication (participation of major

departments) (Papadakis and Barwise 2002;

Papadakis et al. 1998).Theory development has clearly been hindered by

the fragmented nature of the studies reviewed here,

and many have focused solely on a single SDMP

characteristic, often comprehensiveness or rational-

ity. Top management team tenure, education, demo-

graphic diversity, cognitive diversity, polychronicity

and aggressive philosophy all significantly influence

the degree of comprehensiveness. However, the

effects of the CEO on SDMP characteristics are

unclear and, while the effects of TMT demographic

variables on multiple different SDMP characteristics

have been examined (e.g. Papadakis et al. 1998),many of the psychometric variables such as cognitive

diversity have not been subjected to such systematic

scrutiny. Hence, given the limited number of studies

examining TMT variables such as cognitive diversity,

future research may extend this line of enquiry by

studying their effects on other important and fre-

quently studied SDMP characteristics such as

political behaviour, because strong disagreement

concerning strategic issues could conceivably result

in executives forming coalitions and using power to

promote their own perspectives.

The research reviewed in this section raises impor-tant questions concerning whether a focus on an indi-

vidual, such as in the studies of cognitive style, is

appropriate when the SDMP is more often a group

decision process (Hambrick 2007). Therefore, to

complement the laboratory studies on cognitive

style, field-based research focused on teams appears

necessary. Future research examining the influence

of the CEO should focus on small organizations,

where SDM power is more centralized (Milleret al.

1988).

Direct effects of TMT variables on SDMP outcomes

Cognitive style. Hough and ogilvies (2005) labo-

ratory experiment, using the MyersBriggs Type

Indicator to measure executive cognitive style, shows

that cognitive style influences decision quality, deci-

siveness (the number of problems addressed) and

perceived effectiveness (team members perceptions

of one anothers ability to complete tasks). Accord-

ing to Hough and ogilvies (2005) study, executives




9/25

who use both intuition and objective information

make higher-quality decisions, while those who take

the time to make socially acceptable decisions are

indecisive and are perceived to be ineffective.

Other TMT variables. The field experiment using

hypothetical choices developed by Mitchell et al.(2011) found that CEO meta-cognition (reflection on

his or her own thinking) is negatively associated with

erratic SDs (inconsistent judgments that shape the

direction of the firm). It follows, therefore, that

CEOs who reflect on and attempt to control their

cognitive processes are likely to make more consist-

ent SDs. Also, Olsonet al. (2007b) show that cogni-

tive diversity has a negative effect on SD quality and

commitment. Top management teams that disagree

about the strategic goals and priorities of the organi-

zation, therefore, are less likely to be committed to a

SD, and tend to make SDs that fail to achieve thedesired results. However, these effects are reduced

for TMTs who trust one anothers ability and

competence.

As well, several studies have examined the effects

of TMTs and CEOs on the speed of SDM, that is,

how rapidly all aspects of the SDMP are executed,

from the initial consideration of alternative courses

of action to the final commitment to act (Forbes

2005). Top management team polychronicity leads to

speedy SDM (Souitaris and Maestro 2010), and

TMT potency (TMTs perceptions of their ability to

perform tasks effectively) is also positively related toSDM speed (Clark and Maggitti 2012). Contradic-

tory to the studies of CEO demographic variables

that found largely non-significant effects, Wally and

Baum (1994) set up a field-based experiment that

used hypothetical choices, and found that CEO tol-

erance for risk, cognitive ability, use of intuition and

propensity to act are all positively associated with

SDM speed. Furthermore, Forbes (2005) revealed

that younger executives (in their twenties and early

thirties) of new ventures make slower SDs than their

older counterparts. Hence, both TMT and CEO vari-

ables appear to be significant predictors of SD speed,

although the mediating processes that convert

these variables into outcomes (such as speed) remain

unexplained.

TMT summary

This section has revealed the very fragmented nature

of the findings from studies that examine the influ-

ence of the TMT on the SDMP. Theory development

has been hindered because each TMT variable

usually features in no more than one or two studies,

and most studies focus on the effects of TMT vari-

ables on only one SDMP characteristic, often com-

prehensiveness or rationality. These problems are

compounded by the omission of mediating decision

processes, and variance in the level of analysisadopted, with some at the organization level, some at

the individual and others at the decision level. While

studies reviewed here each make a valuable contri-

bution individually, a coherent body of theory has not

developed, and overall implications for theory and

practice remain unclear. The major conceptual and

methodological issues arising from this section are

each now discussed in further detail.

A significant conceptual issue for SDMP research

is the extent to which, relative to other contextual

factors, the TMT influences SDMP characteristics.

Yet few studies examine the influence of TMT vari-ables alongside other categories of contextual vari-

ables. Studies examining the relative influence of

multiple categories of contextual variables either

show that the TMT and CEO have less influence on

SDMP characteristics (Papadakis and Barwise 2002)

or omit the TMT as a category as in the case of

Elbanna and Child 2007b), who examined the rela-

tive influence of the external environment, firm

characteristics and SD-specific characteristics on

rationality. Determining the relative importance of

the TMT for explaining the SDMP compared with

SD-specific characteristics, the external environ-ment, and firm characteristics should be a priority for

future research so as to identify the most pertinent

theories for SDMP research.

The use of demographic variables as proxies for

the underlying cognitive and behavioural traits of the

TMT has attracted criticism, and authors have ques-

tioned how reliably these demographic proxies actu-

ally represent the traits that they are purported to

(Milleret al. 1998; Priem et al. 1999). Despite the

difficulty of obtaining psychometric data from

TMTs, avoiding the use of demographic variables

in future research is recommended. Instead, to attain

greater construct validity and to improve the

explanatory ability of research, direct measures of

the constructs of interest should be obtained.

Another conceptual issue is that no existing

studies examine the moderating effects of the TMT

on the relationship between SDMP characteristics

and SDMP outcomes. This gap is notable, given the

empirical evidence showing how other contextual

variables interact with SDMP characteristics to




10/25

influence SDMP outcomes (Elbanna and Child

2007a). It is highly likely that the effects of SDMP

characteristics on SDMP outcomes are subject to the

influence of TMT variables (Papadakis et al. 2010).

For example, the potential for intuitive SDMPs to

positively influence SDMP outcomes may rest upon

the expertise of the TMT (Khatri and Ng 2000).A final conceptual issue is omission of the medi-

ating SDMP characteristics in studies that specify

direct effects of the TMT on SDMP outcomes. It is

fundamentally important to measure actual mediat-

ing decision processes to fully understand the

causal relationships between contextual vari-

ables, SDMP characteristics and SDMP outcomes.

Studies specifying direct effects of the TMT on

SDMP outcomes do not account for the actual pro-

cesses that convert TMT characteristics into SDMP

outcomes.

Methodological issues include the need for futureresearch to examine the influence of TMTs, rather

than just CEOs, to increase the likelihood of obtain-

ing significant results (Hambrick 2007; Papadakis

and Barwise 2002). Research focused on CEOs

should select samples of small organizations where

SDM power is more centralized. Cognitive style

appears to be an important construct for explaining

SDMP characteristics and outcomes, but existing

studies are laboratory-based and use an individual

level of analysis. Field-based studies focused on the

cognitive styles of teams would shed further light on

its influence.

SD-specific characteristics

Strategic decision-specific characteristics are the

labels and categories that decision-makers attribute

to an SD, based on perceptions of stimuli

(Papadakis et al. 1998). Empirical evidence demon-

strates the significant effects that SD-specific char-

acteristics have on SDMP characteristics, relative to

the effects of other contextual variables (Elbanna

and Child 2007b; Hickson et al. 1986; Papadakis

et al. 1998). In this section, we review studies

of SD-specific characteristics that have featured

decision matter, uncertainty, motive, importance

and time pressure (see Figure 1 and Table 2).

Studies focusing on the direct effects of SD-specific

characteristics on SDMP characteristics are

reviewed first, followed by a review of the literature

examining the moderating effects of SD-specific

characteristics.

Direct effects of SD-specific characteristics on


Decision matter. One of the major conclusions of

the seminal Bradford Studies is that the complexity

and politicality inherent in an SD determine the

process by which it is made (Hickson et al. 2001).

Three types of decision matter are identified

vortex, tractable and familiar that have differing

levels of complexity and politicality (Astley et al.

1982). Each of the three different decision matters

was found to lead to a different decision process,

characterized by varying levels of scrutiny, nego-

tiation, discontinuity, centralization and duration.

However, while the Hicksonet al. (1986) categoriza-

tion conceptualizes politicality as being an anteced-

ent, other studies have modelled politicality as a

characteristic to describe the actual SDMP (e.g.

Dean and Sharfman 1996; Eisenhardt and Bourgeois1988; Elbanna and Child 2007a; Papadakis et al.

1998). To ensure sufficient discriminant validity

between focal constructs, future research should

specify precisely whether politicality is a contextual

antecedent or a characteristic of the SDMP.

SD uncertainty. This diminishes rationality (Dean

and Sharfman 1993) and rule formalization

(Papadakis et al. 1998), but promotes flexibility

(Sharfman and Dean 1997a), politicization and

problem-solving dissension (Papadakis et al. 1998).

Thus, when there is uncertainty concerning theactions that should be taken or the information

required to make a decision (Sonenshein 2007),

decision-makers are less inclined to gather and

analyse information (Dean and Sharfman 1993).

When making uncertain SDs, as well, coalitions are

formed, and bargaining takes place to overcome

internal resistance and disagreement (Papadakis

et al. 1998). However, decision-makers have also

been found to respond to SD uncertainty by being

open to new sources of information (Sharfman and

Dean 1997a).

SD motive. This refers to whether the SD is made in

response to an opportunity or a threat. Papadakis et al.

(1998) discovered that, when the SD is a response to a

threat, the SDMP is characterized by hierarchical

decentralization, as middle management become

involved to a greater extent. However, they found no

other significant effects of SD motive. Fredricksons

(1985) laboratory experiment showed that, while the

MBA sample were more comprehensive when faced




11/25


12/25

with a problem than when faced with an opportunity,

no significant effects were identified in the executive

sample. Similarly, Ashmoset al. (1998) were unable

to find any significant effects of SD motive on the

level of participation in the SDMP (the number of

people involved and the extent of their involve-

ment). Therefore, taken as a whole, these collectiveresults demonstrate only moderate support for SD

motive being a significant influence on SDMP

characteristics.

SD importance. Papadakiset al. (1998) studied the

magnitude of impact of an SD, and discovered that,

when the SD is expected to have profound and per-

vasive implications for the organization and its

operations, decision-makers engage in comprehen-

sive SDMPs, using financial information, involving

management from multiple levels and communicat-

ing across functions. Contrary to this, Dean andSharfman (1993) revealed that decision importance

was not significantly related to rationality. Hence,

whether or not the decision is critical to the organi-

zations future does not appear to determine the

extent of information gathering and analysis, but

does influence comprehensiveness, as well as the use

of financial reporting information, hierarchical

decentralization and lateral communication.

Time pressure. This leads to problem-solving

dissension in the SDMP, and reduces hierarchi-

cal decentralization and lateral communication(Papadakis et al. 1998). Thus, when the SD is under

time pressure, there will be dissent, less involvement

of middle managers and restricted communication

across departments.

Overall, empirical evidence supports the view that

SD-specific characteristics have a significant influ-

ence on the characteristics of the SDMP (Elbanna

and Child 2007b; Papadakis et al. 1998). This has

important implications for studies adopting an

organizational level of analysis, which assumes that

organizations have SDMPs that are consistent

from one decision to the next. Also, the effects of

SD-specific characteristics appear to be nuanced. For

example, the magnitude of impact of the SD is posi-

tively related to comprehensiveness (Papadakiset al.

1998), but the importance of the SD is not related to

rationality (Dean and Sharfman 1993). Such contra-

dictory results highlight the need for the careful

and consistent definition and operationalization of

not only SD-specific characteristics, but also SDMP

characteristics.

Moderating effects of SD-specific characteristics

on the relationship between SDMP characteristics

and SDMP outcomes

Operationalizing rationality as the collection and

analysis of information, Elbanna and Child (2007a)

conclude that the positive relationship between

rationality and SD effectiveness (the extent to which

the decision achieves its objectives and avoids nega-

tive unintended consequences) is stronger for SDs

perceived as crises than for opportunities. Also, the

negative relationship between political behaviour and

SD effectiveness is stronger for SDs perceived as

crises than for opportunities. This evidence suggests

that rational SDMPs work best when faced with a

crisis and that, in such situations, political behaviour

will be even more detrimental to the effectiveness of

the SD. Finally, the authors determine that SD uncer-

tainty moderates the effects of rationality on SDeffectiveness, such that the positive influence of

rationality on SD effectiveness is weaker for low

uncertainty SDs than for high uncertainty SDs.

Therefore, when decision-makers are uncertain

about the information required and the likely out-

comes, rational processes are more likely to produce

effective SDs.

SD-specific characteristics summary

Empirical evidence suggests that SD-specific charac-

teristics may be one of the most significant influences

on SDMP characteristics, relative to other contextualvariables (Elbanna and Child 2007b; Hickson

et al. 1986; Papadakis et al. 1998). Despite this,

SD-specific characteristics are featured in only a

handful of studies, and only one empirical study spe-

cifically addresses their moderating influence on the

relationship between SD-specific characteristics and

outcomes. Hence, it is doubtful that the findings of

the studies reviewed here are generalizable. Also, it

remains unclear why, for example, the magnitude of

impact from an SD and SD importance each has

different implications for comprehensiveness and

rationality (Dean and Sharfman 1993; Papadakis

et al. 1998).

It is also apparent that many of the SD-specific

characteristics used in the studies reviewed here

suffer from low reliabilities, and several studies have

reported Cronbach alphas of between 0.54 and 0.63

(e.g. Elbanna and Child 2007a; Dean and Sharfman

1993; Papadakis et al. 1998). Low reliabilities indi-

cate high levels of random error in the measures,

which reduces the probability of detecting significant




13/25

relationships (Bagozzi 1994). Therefore, research

focusing on SD-specific characteristics will benefit

from measure development work to improve their

reliabilities.

Given the significant interactions detected in

Elbanna and Childs (2007a) study and the rela-

tive importance of SD-specific characteristics asantecedents of SDMP characteristics, a systematic

programme of research examining SD-specific char-

acteristics is warranted. Such research should model

SD-specific characteristics as antecedents of differ-

ent SDMP characteristics, and as moderators of the

relationships between SDMP characteristics and

SDMP outcomes, which would significantly help to

develop a more coherent body of theory.

The external environment

The external environment is the third category of

contextual variables, and in this section we first

review those empirical studies that examine the

direct effects of the external environment on SDMP

characteristics. Second, we review studies that test

the direct effects on SDMP outcomes and, finally, we

review those that feature the moderating effects of

the external environment on the relationship between

SDMP characteristics and SDMP outcomes. The

dimensions of the external environment reviewed in

this section include environmental hostility, velocity,

dynamism, instability, munificence and uncertainty(see Figure 1 and Table 3).

Direct effects of the external environment on


Hostile environments. Hostile environments are

threatening and dangerous, making it difficult to

stay afloat (Miller and Friesen 1983), and studies

have produced confounding results. Dean and

Sharfman (1993) and Sharfman and Dean (1997a)

conclude that in such environments, decision-

makers are less likely to gather and analyse infor-

mation, and are less open to new information.

However, Papadakis et al. (1998) were unable to

detect significant effects of environmental hostility

on any of the SDMP characteristics. Aside from

being contradictory, the results of these studies are

unexpected, given that in such environments the

implications of making an ineffective SD are severe

and, hence, more rational or comprehensive SDMPs

would be expected.

High-velocity, dynamic and unstable environ-

ments. High-velocity environments are character-

ized by rapid and discontinuous change in demand,

competitors, technology and/or regulation, such that

information is often inaccurate, unavailable, or

obsolete (Bourgeois and Eisenhardt 1988, p. 816).

Eisenhardt (1989) uses eight case studies of firms inthe high-velocity micro-computer industry to high-

light how organizations use real-time information,

and simultaneously consider multiple SD alterna-

tives in such environments. Dynamic environments

are very similar to high-velocity environments

(Baum and Wally 2003) and are characterized by a

highly unpredictable and unstable rate of change

and high levels of uncertainty about the state of

context, the meansends relationships, and/or the

outcomes of the actions (Mitchell et al. 2011, pp.

687688). However, Papadakis et al. (1998) is the

only empirical study to examine the direct effects ofenvironmental dynamism on SDMP characteristics,

and this study produced no significant effects. Envi-

ronmental instability, defined as the extent to which

market demand and technology are rapidly changing

in a given industry (Dean and Sharfman 1996, p.

376), has been found to cause TMT members to

exhibit more agreement about the comprehensive-

ness (the actual process was not reported) of the

SDMP than in stable environments (Iaquinto and

Fredrickson 1997). Hence, in unstable environments,

executives are more likely to focus on the SDMP and

not become distracted by disagreements and debates.Based on the studies reviewed here, it may be true

that the external environment is a less significant

influence on SDMP characteristics compared with

the other categories of contextual variables. Cer-

tainly, the lack of significant findings in Papadakis

et al. (1998) for both environmental hostility and

dynamism raises doubts over the veracity of the envi-

ronmental determinism perspective (Aldrich 1979;

Hannan and Freeman 1977). Also, the work of

Elbanna and Child (2007b) challenges the signifi-

cance of the external environment as a means for

explaining the characteristics of the SDMP; the

authors determine that SD-specific characteristics

and firm characteristics are more significant predic-

tors of rationality than the external environment is.

Direct effects of the external environment on

SDMP outcomes

Environmental dynamism. Mitchell et al. (2011)

use a field experiment with hypothetical choices to




14/25


15/25


16/25

determine that environmental dynamism causes

CEOs to make fewer erratic SDs, that environmental

hostility causes CEOs to make more erratic SDs, and

that dynamism and hostility interact to cause CEOs

to make fewer erratic SDs. Thus, not only do envi-

ronmental dynamism and hostility exert different

effects, they interact to influence the outcomes of theSDMP. Baum and Wally (2003) also use a field

experiment with hypothetical choices, and find that

environmental dynamism and environmental munifi-

cence (capacity to support growth) both cause CEOs

to make faster SDs. Both the Mitchell et al. (2011)

and Baum and Wally (2003) studies omit mediating

decision processes, which leaves unexplained the

actual decision processes that mediate the effects of

the external environment on decision speed and

whether the SD is erratic.

Moderating effects of the external environment on

the relationship between SDMP characteristics and

SDMP outcomes

High-velocity, dynamic and unstable environments.

A common theme in the literature has been to iden-

tify interactions between the external environment

and SDMP characteristics, and their influence on

SDMP outcomes. Several studies have shown that in

high-velocity or dynamic environments, rational and

comprehensive SDMPs are positively associated

with organizational performance (Bourgeois and

Eisenhardt 1988; Eisenhardt 1989; Priem et al.1995). The basic tenet of these studies is that the

uncertainty inherent in such environments can be

mitigated by comprehensively gathering and analys-

ing information, and developing and evaluating mul-

tiple different decision options. However, there is

also empirical evidence to suggest that the opposite

is true, and that in such environments rational or

comprehensive approaches to SDM are of little

value, because information is unavailable, is incom-

plete or becomes obsolete rapidly. The field-based

experiments by Fredrickson and colleagues that use

hypothetical scenarios with executives show that, in

unstable environments, comprehensive SDMPs have

a negative effect on organizational performance.

These findings are also supported by Hough and

White (2003), who use a laboratory simulation to

show how rational SDMPs have no effect on decision

quality in dynamic environments. It should also be

noted, however, that no significant interactions were

found between rationality and environmental insta-

bility (Dean and Sharfman 1996) and rationality and

environmental uncertainty (Elbanna and Child

2007a).

There is a clear tension in thefindings of Eisenhardt

and colleagues studies, as compared with those of

Fredrickson and colleagues. This may be explained

by the sharp and discontinuous change (Bourgeois

and Eisenhardt 1988, p. 816) that distinguishes high-velocity environments from the unstable environ-

ments studied by Fredrickson and colleagues.Also, in

research that has revealed positive effects of SDMP

comprehensiveness in high-velocity environments, a

common theme is the importance of SD speed.

Eisenhardt (1989) highlights the ways in which

decision-makers consider multiple decision options

simultaneously, using real-time information to make

speedy SDs and to achieve superior organizational

performance. Similarly, Judge and Miller (1991) also

determine SD speed to be positively related to

organizational performance though only in high-velocity environments, and not in medium- or low-

velocity environments. A tentative conclusion may be

reached that an organizations ability to engage in

SDMPs that are both comprehensive and fast leads to

high levels of organizational performance in rapidly

changing environments.

Environmental munificence. Other studies have

focused on how environmental munificence interacts

with rationality to influence SDMP outcomes, and

again no clear consensus exists. While Goll and

Rasheed (2005) show that rationality leads to highlevels of organizational performance in munificent

environments, Elbanna and Child (2007a) report that

rationality is most likely to result in an effective SD

in environments with low levels of munificence

where few opportunities for growth exist.

Empirical evidence also suggests that interactions

between the external environment and SDMP char-

acteristics may be even more complex than most

existing research allows for. Goll and Rasheed

(1997) show that not only does rationality benefit

organizational performance in dynamic and in

munificent environments, but the positive effects of

rationality on performance are strongest in environ-

ments high in both dynamism and munificence.

Environmental uncertainty. Atuahene-Gima and Li

(2004) adopt a sophisticated approach to modelling

the external environment to examine two types of

uncertainty: demand uncertainty and technology

uncertainty. Their results indicate that comprehen-

sive SDMPs benefit new product performance and




17/25

quality in the presence of demand uncertainty,

because organizations are able to mitigate such

uncertainty by searching for and analysing informa-

tion relating to customer demand and preferences.

However, in the presence of technology uncertainty,

comprehensive SDMPs are of little benefit, as

information relating to technology uncertaintyis dense, highly equivocal and difficult to collect

(Atuahene-Gima and Li 2004). These findings point

to a more nuanced interaction between SDMP char-

acteristics and the external environment than most

studies have modelled.

Other interaction effects involving the external

environment. Studies have also examined how the

external environment interacts with other SDMP

characteristics such as political behaviour and intui-

tion. Bourgeois and Eisenhardt (1988) reveal that

political behaviour causes poor organizational per-formance in high-velocity environments, because it

distracts executives and causes delays (Eisenhardt

and Bourgeois 1988). Also, Khatri and Ng (2000)

studied intuitive synthesis (reliance on judgment,

experience, and gut feelings) in the SDMP and

found a positive relationship with organizational

performance in unstable environments, but they

observed a negative relationship in stable environ-

ments. These results suggest that intuitive synthesis

can help to overcome the limitations of rational and

comprehensive SDMPs in unstable environments,

perhaps because of its speed and ability to synthesizeinformation, but intuitive synthesis needs to be used

cautiously in stable environments (Khatri and Ng

2000).

External environment summary

Theory development has been hindered by the

absence of a single approach to conceptualizing and

measuring the external environment (Sharfman and

Dean 1991). Studies have used environmental veloc-

ity, instability and dynamism to capture the extent to

which the external environment is subject to rapid

and unpredictable change, which has produced

conflicting results concerning the effects of com-

prehensiveness on performance. Similarly, studies

examining the implications of rationality in munifi-

cent environments have produced contradictory

results. Also, relative to other contextual variables,

the external environment appears to have limited

direct effects on SDMP characteristics (Elbanna and

Child 2007b; Papadakis et al. 1998), and its influ-

ence seems to be primarily a moderating one, influ-

encing the relationship between SDMP charac-

teristics and outcomes. In part, the contradictory

findings of the studies reviewed in this section

may be caused by some studies adopting an

organizational level of analysis and others focusing

on the decision level. Matters have been complicatedby differences in the SDMP characteristics used,

with some examining rationality and others compre-

hensiveness, and with the varying way in which these

constructs are operationalized.

There is a clear need to reconcile the conflicting

results of the Eisenhardt and Fredrickson studies,

concerning the relationship between comprehen-

siveness and organizational performance. Such

research should focus on the implications of com-

prehensiveness and SD speed combined, under dif-

ferent environmental conditions, because, despite

the apparent importance of speed in fast-changingenvironments (Eisenhardt 1989), most studies omit

it. Also, future research may try to reconcile the

inconsistencies of other studies, such as those con-

cerning the implications of rationality in munificent

environments (Elbanna and Child 2007a; Goll and

Rasheed 2005) through measurement of multiple

different dimensions of the external environment

within a single study. This would enable the detec-

tion of three-way interactions shown to be impor-

tant by Goll and Rasheed (1997). Furthermore,

future research may also benefit from careful atten-

tion to modelling the external environment. Theapproach adopted by Atuahene-Gima and Li (2004)

is rare, in that it examines the effects of alternative

types of environmental uncertainty (technology

and demand) on the relationship between com-

prehensiveness and SDMP outcomes. Therefore, a

focus on the different types of uncertainty, instabil-

ity, dynamism and velocity, such as customers,

competitors and technology, may provide fresh

insights.

Firm characteristics

In this section, we review the literature examining the

direct effects of firm characteristics on SDMP char-

acteristics and outcomes. Also, we review studies

that examine the moderating effects of firm charac-

teristics on the relationship between SDMP charac-

teristics and outcomes. Firm characteristics featured

in this section include power centralization, struc-

ture, size, performance, slack resources, external




18/25

control, corporate control and planning formality

(see Figure 1 and Table 4).

Direct effects of firm characteristics on


Power centralization. Power centralized in thehands of the CEO leads to political behaviour

(Bourgeois and Eisenhardt 1988; Eisenhardt and

Bourgeois 1988). Hence, when the CEO does not

distribute SDM power among members of the TMT,

team members form coalitions, pursue their own

agendas and, overall, act in a way that is destructive

to SDs (Bourgeois and Eisenhardt 1988).

Structure. Structure influences participation and

rationality in the SDMP. Ashmos et al. (1998) show

that organizations with fewer rules and standardized

operating procedures have greater participation ofmultiple different internal stakeholders in the SDMP,

whereas, in organizations with formalized rules, the

SDMP is consistent and formulaic, with the same

small number of individuals involved. Furthermore,

Miller (1987) concludes that organizations with for-

malized systems and processes that use task forces

and committees will also favour rational SDMPs. For

the study, Miller (1987) adopts an organizational

level of analysis, and defines rationality as analysis,

systematic scanning of the environment and having

explicit strategies.

Size. Size is commonly operationalized as the

number of full-time employees, and empirical evi-

dence shows that certain SDMP characteristics vary

according to the size of the organization. Large

organizations are more comprehensive in their

SDMPs than small organizations (Fredrickson and

Iaquinto 1989; Papadakis et al. 1998), although the

effects of size on other SDMP characteristics are less

clear. Iaquinto and Fredrickson (1997) report that as

firms increase in size, TMT agreement about the

comprehensiveness of the SDMP diminishes which

implies that levels of debate and disagreement

regarding the SDMP will increase in large organiza-

tions. However, Papadakis et al. (1998) found no

association between size and politicization. Simi-

larly, Dean and Sharfman (1993) were unable to

detect any relationship between size and rationality.

Overall, these findings suggest that, when firms

increase in size, so too does the comprehensiveness

of their SDMPs, but the implications of size for

rationality and political behaviour remain unclear.

Performance. Another line of empirical enquiry

has been to examine the effects of organizational

performance on SDMP characteristics, although

confounding results have emerged. Papadakis et al.

(1998) used two measures of organizational perfor-

mance and found that organizations generating a

high return on assets tend to have SDMPs that arecharacterized by comprehensiveness, the use of

financial reporting information and hierarchical

decentralization, whereas the SDMPs of organiza-

tions experiencing profit growth are more political.

Papadakis et al. (1998) suggest that high levels of

organizational performance endow organizations

with resources to invest in comprehensive and

decentralized SDMPs that make greater use of

financial reporting information. However, perfor-

mance also provokes politics and conflict over how

best to use these resources. Fredricksons (1985)

laboratory experiment was unable to reach a con-clusion concerning the effects of performance on

comprehensiveness. MBA students were more com-

prehensive when organizational performance was

poor compared with when performance was excel-

lent, whereas the sample of executives exhibited no

differences according to performance. Also,

Iaquinto and Fredrickson (1997) found no signifi-

cant effect of performance on TMT agreement

about comprehensiveness. Finally, Sharfman and

Dean (1997a) investigated the effects of slack

resources (resources intentionally kept beyond those

needed to meet operational commitments) on theflexibility of the SDMP. Organizations with high

levels of slack resources were found to be open

to new ideas, to new sources of information, and

to decision-makers adopting new roles in the

SDMP.

Other firm characteristics. Other firm characteris-

tics affecting SDMP characteristics include the

degree of external control, type of corporate control

and planning formality. When the SDMP is subject to

the influence of individuals external to the organiza-

tion, the SDMP becomes less rational (Dean and

Sharfman 1993), thus highlighting the importance of

managerial discretion for rationality. Papadakiset al.

(1998) examined the type of corporate control and

found that the SDMPs of state-controlled organiza-

tions are comprehensive and political, whereas pri-

vately owned organizations are less comprehensive,

and rely less on financial reporting and formalized

rules. Also, organizations with formal planning

systems have greater lateral communication, and are




19/25


20/25

more comprehensive and political (Papadakis et al.

1998).

From this section of the review it is possible to

reach a number of tentative conclusions: (1) Power

centralization invokes political behaviour; (2) As

organizations increase in size, their SDMPs become

more comprehensive; (3) Organization structureinfluences rationality, as well as who participates in

the SDMP, and; (4) Ownership, external control

and formal planning influence comprehensiveness,

rationality and political behaviour.

One particularly ambiguous area concerns the

effects of organizational performance on the SDMP,

with Papadakis et al. (1998) concluding that return

on assets and growth in profits each affect diff-

erent SDMP characteristics. However, adopting

measures of performance such as return on assets

and growth in profits may not reliably reflect the

actual levels of resources available for use by theorganization. For example, an organization could

be highly profitable, but also highly geared, with

significant annual cash outflows to repay its debt.

Such cash outflows would not be reflected in meas-

ures of performance such as profit growth, but

would result in the organizations being resource

constrained. Furthermore, performance is unlikely

to be a valid proxy for other types of slack

resources (such as human resource slack) (Voss

et al. 2008).

Direct effects of firm characteristics on

SDMP outcomes

Studies have examined the direct effects of firm char-

acteristics on SD speed. While Eisenhardts (1989)

case studies do not reveal any link between power

centralization and SD speed, Wally and Baum (1994)

and Baum and Wally (2003) both demonstrated that,

when authority for SDM is centralized, SD speed

increases.

Moderating effects of firm characteristics on therelationship between SDMP characteristics and

SDMP outcomes

Covinet al. (2001) identified an interaction between

intuitive and technocratic SDMPs, organization

structure and the external environment, which influ-

ences organization performance. In high-technology

environments, intuitive SDMPs lead to high levels of

organization performance among firms with organic

structures (characterized by open communication

and few rules and procedures). Covin et al. (2001)

also determined that in low-technology environ-

ments, technocratic SDMPs (using systematic

and quantitative analysis) lead to high levels of

organizational performance among firms with

organic structures. Thus, organizations need to

ensure that their SDMPs and organizational struc-tures are configured to best fit their external environ-

ments (Covin et al. 2001).

Elbanna and Child (2007a) found that the posi-

tive influence of rationality on SD effectiveness is

moderated by organizational performance (using a

perceptual measure of both financial and non-

financial organizational performance), such that in

high-performing organizations the influence is

weaker. Additionally, the negative influences of

intuition and political behaviour on SD effective-

ness are weaker for companies with high perfor-

mance. While there is scant other empiricalevidence with which to compare these findings, it

does highlight a view that the effects of SDMP

characteristics on SD effectiveness are subject to

the influence of firm characteristics.

Firm characteristics summary

Firm characteristics exert a significant influence on

SDMP characteristics relative to other contextual

variables (Elbanna and Child 2007b). For instance,

power centralization affects political behaviour,

structure influences rationality and participation, andsize is significantly related to comprehensiveness.

External control, corporate control and formal plan-

ning also influence rationality, comprehensiveness

and political behaviour. However, organizational per-

formance has produced confounding results, with

return on assets and growth in profits each influenc-

ing different SDMP characteristics. Careful theory

development is required to disentangle the differing

effects of alternative measures of organizational per-

formance (Papadakis et al. 1998). Because perfor-

mance has been used as a proxy for slack resources,

future research should consider using direct meas-

ures of slack resources to examine its effects on

SDMP characteristics, as well as its moderating

influence on the relationship between SDMP charac-

teristics and SDMP outcomes. Additionally, future

research should examine the moderating effects

of other firm characteristics and, in particular,

should consider the possibility of three-way interac-

tions among firm characteristics and contextual

variables such as the external environment that




21/25

are highlighted by Covin et al. (2001) as hav-

ing significant implications for organizational

performance.

Despite calls in the literature for such research

(e.g. Eisenhardt 1989; Molloy and Schwenk 1995), it

is evident that no studies have examined firm capa-

bilities such as real-time information systems. Theabsence of such research is particularly noteworthy,

given that the ability to make comprehensive and fast

SDs is vital to achieve superior levels of performance

in high-velocity environments (Eisenhardt 1989).

Such capabilities may rest upon the information

systems and technologies of the firm.

Conclusions

This review has highlighted the underlying themes,

issues, tensions and debates in the SDMP literatureregarding the direct and moderating influences of

context. In this section, we discuss the priorities for

future theory development. These include the need

for fewer constructs and for more careful modelling,

as well as the importance of examining multi-

theoretic models and the moderating effects of con-

textual variables. We also address the methodological

implications, which include the importance of large

samples and multivariate analysis, measurement reli-

ability, issues surrounding levels of analysis, and

qualitative research designs.

Priorities for future theory development

The need for fewer constructs and for more careful

modelling. An incremental approach to theory

building has resulted in a proliferation of constructs,

especially SDMP characteristics, and there is con-

sensus in the literature that too much invention of

language (Bower 1997, p. 27) has hindered theory

development (Papadakis and Barwise 1997). This

review has identified significant differences in the

definitions and operationalizations of rationality and

comprehensiveness, and they are frequently used

interchangeably (Goll and Rasheed 1997, 2005;

Papadakis and Barwise 2002; Papadakiset al. 1998;

Priemet al. 1995). This is despite evidence that con-

textual variables such as SD importance and firm size

affect rationality and comprehensiveness differently.

To address problems with the overabundance of

SDMP characteristics, Elbanna (2006) and Elbanna

and Child (2007a) highlight how rationality, compre-

hensiveness, intuition and political behaviour are a

parsimonious set of constructs that adequately repre-

sent two perspectives underpinning the SDMP litera-

ture, namely synoptic-formalism and incremen-

talism. Synoptic formalism stresses formal analysis

in the SDMP, and incrementalism views SDMPs as

subject to behavioural influences such as intuition

and politics (Eisenhardt and Zbaracki 1992;

Fredrickson and Mitchell 1984; Hart 1992; Hitt andTyler 1991). Elbanna (2006) and Elbanna and Child

(2007a) suggest that rationality and comprehensive-

ness are constructs that represent the synoptic formal

perspective, and political behaviour and intuition are

constructs that embody the incremental perspective.

Certainly, a common terminology and consistency in

the variables modelled would support the develop-

ment of a more coherent body of theory (Papadakis

and Barwise 1997). Furthermore, future research

should precisely define and operationalize rationality

and comprehensiveness, and should have a clear

theoretical rationale for including one or the other.Including both in one study would enable researchers

to discern the relative influence of context on each of

them.

Lack of replication. Lack of replication has hin-

dered the coherent development of theory. Future

research should seek to replicate, or at least control

for, some of the significant variables and relation-

ships identified in this review. For example, decision

speed has significant implications for perfor-

mance (Eisenhardt 1989; Judge and Miller 1991);

despite this, most studies omit it. Similarly, studiesattempting to explain SDMP outcomes under certain

environmental conditions have focused on com-

prehensiveness or rationality. However, there is

considerable empirical evidence that SDMPs are

multi-dimensional, and SDMP outcomes are subject

to the influence of SDMP characteristics other than

just rationality, such as political behaviour (Dean and

Sharfman 1996). Closer attention to control variables

would lead to greater implicit replication.

Direct measures. Using demographics as proxies

for the underlying psychometric characteristics of

the TMT has received strong criticism, and Priem

et al. (1999) suggest that this approach sacrifices

construct validity for measurement reliability, and

forsakes explanation for prediction. Similarly, this

review identifies organizational performance being

used as a proxy for slack resources. To improve

methodological rigour and to develop theory with

greater explanatory power, future research should

use direct measures of contextual variables.




22/25

Mediating decision processes. Researchers should

include measures for actual decision processes. For

example, examining the effects of cognitive style on

decision outcomes neglects the mediating role of the

actual decision process. Because a decision-maker

has a rational cognitive style, it does not necessarily

follow that the actual SDMP will be rational; othercontextual variables may have a more significant

effect in determining the decision process (Bakken

and Haerem 2011). It is fundamentally important to

measure actual mediating decision processes to fully

understand the causal relationships between context,

SDMP process characteristics and SDMP outcomes.

The need for multi-theoretic models

Some authors (e.g. Elbanna and Child 2007b; Hitt

and Tyler 1991; Papadakis et al. 1998) have associ-

ated each of the four categories of contextual vari-

ables with a theoretical perspective. For example, the

TMT is associated with the upper echelons perspec-

tive (Hambrick and Mason 1984), and the external

environment is associated with the environmental

determinism perspective (Aldrich 1979; Hannan and

Freeman 1977). By including contextual variables

from each of the four categories, researchers can

compare the overall impact and predictive power of

each of these theoretical perspectives for explaining

variance in SDMP characteristics. While most

studies have focused on one category of contextual

variable, there is no theoretical rationale, as yet, topropose that any category is more, or less, significant

than any other (Rajagopalan et al. 1993).

The moderating role of contextual variables

With the exception of the external environment, there

is scant empirical research examining the moderating

effects of other contextual variables on the relation-

ship between SDMP characteristics and outcomes,

and even research examining the moderating effects

of the external environment has produced conflicting

results. Furthermore, TMT variables are yet to be

featured as moderators. Models that incorporate bothSDMP characteristics and contextual variables as

moderators can predict more variance in outcome

variables, because the limited empirical evidence

that exists has shown such interactions to be highly

significant (Elbanna and Child 2007a).

Methodological implications

Large samples and multivariate analysis. Most of

the managerially relevant SDM research is based on

a fairly small number of cases studied in depth

(Papadakis and Barwise 1997, p. 296). The insights

provided by Bourgeois and Eisenhardt (1988),

Eisenhardt (1989), Eisenhardt and Bourgeois (1988),

Dean and Sharfman (1993, 1996), Papadakis and

Barwise (2002) and Papadakiset al. (1998) all draw

upon samples with 70 or fewer SDs from 38 or fewerorganizations. Therefore, to improve the external

validity and generalizability of SDMP research, there

is a need for large-sample field research (Papadakis

et al. 2010; Rajagopalan et al. 1993). If multiple

contextual variables, SDMP characteristics and

SDMP outcomes are to be modelled, larger sample

sizes will be essential to ensure that the ratio of cases

to predictor variables is sufficient. Multivariate

analysis comprises a set of statistical analysis tech-

niques well suited to such analysis.

Measurement reliability. Measures of SD-specificcharacteristics have suffered from low levels of inter-

nal consistency, which indicates high levels of

random error. Unreliable measures make detecting

significant relationships less likely (Bagozzi 1994).

Given the empirical evidence suggesting that

SD-specific characteristics are potentially one of the

most significant contextual influences on the

SDMP, future research should seek to improve their

reliabilities.

Levels of analysis. Variance in the level of analysis

adopted has hindered theory development and makescomparisons of findings difficult. While some studies

do focus on individual decisions, a large number

adopt an organizational level of analysis, assuming

that organizations have consistent SDMPs, despite

considerable empirical evidence showing that the

SDMP varies according to the individual SD being

made (Elbanna and Child 2007b; Hickson et al.

1986; Papadakis et al. 1998). A focus on individual

decisions is therefore recommended. Furthermore,

there is a lack of significant results from studies

examining the effects of CEO characteristics on the

SDMP. Unless the focus of research is small organi-

zations, it appears likely that the characteristics of the

TMT as a whole will have more significant effects on

the SDMP, rather than any one individual.

Qualitative research designs. To complement large

sample studies and to examine emerging con-

cepts such as non-conscious cognitive processes

(Hodgkinson and Healey 2011), SDMP researchers

could adopt methods commonly used by scholars in




23/25

other domains particularly cognitive task analysis

(CTA) and critical incident technique (CIT) to

probe incidents and uncover contextual influences on

the SDMP. For example, techniques such as CIT and

CTA can lead to a better understanding of the con-

textual variables that contribute to intuitive hits and

misses (Hodgkinson and Sadler-Smith 2011, p. 60).These types of techniques could be conducted with

multiple members of the TMT, creating potential to

uncover different perspectives from those involved in

the SDMP.

Summary

This review has identified four categories of con-

textual variables (TMT, SD-specific characteristics,

the external environment and firm characteristics)

that influence the SDMP, and provided an in-depth

analysis of the underlying themes, issues, tensionsand debates in the literature. To develop theory in

this important area, future research should take

greater care with modelling, and include mediating

decision processes, pay close attention to control

variables and use direct measures of contextual

variables. Future studies could also include contex-

tual variables from each of the four categories to

determine their relative influence, and also model

contextual variables as moderators of the SDMP

characteristicsoutcomes relationship, as well as

antecedents of the SDMP. The important methodo-

logical implications arising from this review high-light the importance that future research should use

large sample sizes and multivariate techniques,

improve measurement reliability, adopt a decisional

level of analysis and complement quantitative

studies with qualitative research.

References

Aldrich, H.E. (1979). Organizations and Environments.

Englewood Cliffs, NJ: Prentice-Hall.

Ashmos, D.P., Duchon, D. and McDaniel, R.R. Jr (1998).

Participation in strategic decision making: the role oforganizational predisposition and issue interpretation.

Decision Sciences, 29, pp. 2551.

Astley, W.G., Axelsson, R., Butler, R.J., Hickson, D.J. and

Wilson, D.C. (1982). Complexity and cleavage: dual

explanations of strategic decision-making. Journal of

Management Studies, 19, pp. 357375.

Atuahene-Gima, K. and

environment and sdmprocess review 2014

Documents