You are on page 1of 15

JOURNAL OF MULTI-CRITERIA DECISION ANALYSIS

J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)


Published online in Wiley InterScience
(www.interscience.wiley.com). DOI: 10.1002/mcda.437

Application of Multiple Criteria Decision Analysis


in the New Zealand Agricultural Industry
A. E. DOOLEY, D. C. SMEATON, G.W. SHEATH and S. F. LEDGARD
AgResearch Ltd, Hamilton, New Zealand

ABSTRACT
Agricultural decisions typically involve multiple criteria, some of which are subjective. Business, environmental, and
lifestyle criteria are all important criteria in these decisions. These criteria can be difficult to trade-off using traditional
methods. Multiple criteria decision analysis (MCDA) provides a formal, quantitative means of evaluating agricultural
decisions taking all these factors into account. MCDA for agricultural decisions was evaluated using three applied
case studies (use of new processing technology, selection of beef policies, and selection of farm systems in an
environmentally sensitive catchment). The case studies all differed in their problem types and decision-maker
requirements. A multi-attribute value theory approach was used for all the cases. This approach was selected using a
descriptive framework which took the method limitations, problem attributes, and decision-maker requirements into
account. The differences between the case studies, the application and the implementation of MCDA, the overall
success of the process, and the potential use of MCDA in agriculture going forward are discussed. While MCDA was
used to help identify the best decision, the main benefits that the decision makers identified included: learning about
the decision, a better understanding of their own and others’ perspectives, a structured way to work through the
decision, a means to explain the decision, and stimulation of discussion and sharing of ideas. These benefits were
particularly important for the group decisions. Participants were not overly concerned with the ranking accuracy.
Problems in implementation included an initial lack of commitment to the process, understanding of the process and
decision, and ownership of the decision. The limited time decision makers had available contributed to this. The
majority of the decision makers liked the MCDA process for these strategic decisions. The quantitative approach and
the graphic presentation appealed to them. Hence, simple MCDA approaches may be as effective as more complex
ones, and can deliver many of the benefits particularly where the time is limited. There has been interest in the use of
MCDA going forward. Copyright r 2010 John Wiley & Sons, Ltd.

KEY WORDS: multiple criteria decision analysis; agriculture; environment; New Zealand; decision making

1. INTRODUCTION future. Agricultural businesses will have to remain


viable both environmentally and economically,
Agriculture, forestry, and horticulture remain while continuing to operate in a business environ-
significant industries in the New Zealand’s econ- ment that is likely to require increasingly higher
omy. These industries operate without subsidies. standards. As farming is also often a lifestyle
The majority of the products are exported, earning choice, lifestyle factors also need to be considered
60% of the country’s export revenue and con- in many agricultural business decisions.
tributing 18% to GDP. New Zealand is known for Multiple criteria decision analysis (MCDA) has
its’ ‘clean, green’ pasture-based farm production the potential to be useful in agricultural and
systems and high-quality products: attributes that environmental decision making to help trade-off
are important to our customers. The increasing the economic, environmental, and social aspects that
interest in sustainability, climate change and need to be considered in making strategic decisions.
environmental impacts, both internationally and MCDA is a formal quantitative approach for
nationally will impact on agriculture more in the exploring decisions and assisting with decision
making where there are multiple, conflicting goals
measured in incommensurate units (RAC, 1992;
*Correspondence to: AgResearch Ltd, Ruakura Beinat, 2001; Belton and Stewart, 2002). Other
Research Centre, Private Bag 3123, Hamilton, advantages can include: making a decision more
New Zealand. E-mail: liz.dooley@agresearch.co.nz transparent to others, providing a focus for

Copyright r 2010 John Wiley & Sons, Ltd.


40 A. E. DOOLEY ET AL.

discussion, providing a means of problem Three diverse agricultural case studies involving
structuring and working through the information, strategic decisions at a business level were evaluated
and breaking the decision down so that people (Table I). While all the three case studies were
better understand the decision both from their from the agricultural sector, they were very
own and from others’ perspectives (Belton and different in their problem types and decision-
Stewart, 2002). maker requirements. A descriptive framework
Traditionally, multiple objectives in agriculture which took into account the problem attributes,
have been traded-off formally using cost-benefit and method and decision-maker requirements was
analysis, but valuation of environmental factors developed to help select the most appropriate
can prove difficult and controversial (RAC, 1992; MCDA methods to use in the case studies. The
Roy, 1999). Consideration of these environmental three case studies were:
issues in agricultural policy decisions has become
increasingly important over the last decade  An agribusiness decision: use of new technology
(Beinat, 2001) and ways to incorporate them are in an agricultural processing company.
needed. Farm business level strategic decisions  Farm business decisions: farmers selecting the
often combine a formal economic analysis, with most suitable beef cattle policy.
less formal consideration of other factors  Farm business decisions: farmers in an envir-
(Shadbolt and Martin, 2005). While MCDA has onmentally sensitive catchment evaluating farm
been used at all levels of agricultural decision systems that could be implemented in order to
making, from farm-level to agricultural policy reduce environmental impacts.
decisions, it has rarely been used for agricultural
and environmental decision making in Australia This paper will first describe the three case
and New Zealand (RAC, 1992). studies, followed by the MCDA method selection
There is considerable literature on MCDA and evaluation processes. The case study
techniques, particularly the various aggregation implementation and process evaluation will then
procedures, and reports on results obtained be discussed followed by a section on the potential
using hypothetical data or non-typical decision for MCDA in future agricultural decision making
makers. However, there is limited literature on in New Zealand.
applications of MCDA, how to choose between
approaches or techniques, or why a particular
approach was chosen (Belton & Stewart, 2002; 2. DESCRIPTION OF THE THREE CASE
Belton, 2001; French, 2000, 1998). The scarcity STUDIES
of literature on real MCDA applications is
suggested by Roy (1999) to be because of the MCDA applications reported in the literature are
considerable work involved in describing a real often policy decisions with large numbers of
decision process with all its complications, and stakeholders or stakeholder groups. In contrast,
results that lack the characteristics required for our case study decisions were strategic decisions
many scientific journals. Eden (1995) and made at a business level, although one decision
Montibeller (2007) argue the need for more was looking at the responses of individual farmers
action research to capture the often complex to environmental policy changes. The case studies
reality of decisions and the context in which they selected largely reflect the nature of farming
are being made. Eden (1995) noted that real life businesses in New Zealand.
decisions differ from those in which students or The participants in the case studies were already
others act as decision makers: in reality the working with AgResearch on other projects which
decision makers have to invest time and money required a decision to be made (Note: AgResearch
into the decision, accept the risk and live with the Ltd. is New Zealand’s largest government-owned
consequences, and work together to implement the research organization, delivering agricultural
decision. research from ‘pasture to plate’). The participating
In this paper we will describe an action research businesses agreed to the inclusion of an MCDA
approach to evaluate the usefulness of MCDA in process to assist with the decisions they were
agricultural decision making. Learnings from the considering. Using real cases ensured that our
implementation of MCDA in an agricultural results reflected experiences and feedback from
decision-making context will be identified. people in genuine decision-making situations

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda
APPLICATION OF MULTIPLE CRITERIA DECISION ANALYSIS 41

Table I. Description of the case studies


Agribusiness (AB) Farm business—Beef (FB) Farm business—Taupo (AE)

Problem Identify best strategy for objective Identify the best beef policy (beef Identify the best farm system for
carcass measurement technology use cow farm and a finishing cattle farm) Lake Taupo catchment properties
People Operations managers (OM) for Mentor group assisted in identify- 4 Maori groups (trustee, manager
involved procurement, processing, and mar- ing alternatives (2 meetings). Man- and/or consultant). 3 individual
keting. Senior managers (SM) had ager (cows) and farm owners farmers (couples or individuals). 4
input at an initial and final group (finishing) completed the analysis researchers
meetings 2 Environment Waikato staff
Context Part of another project. OM were Agreed to include MCDA in an- Part of the research with farmers in
told by SM to participate and come other project with farmers wanting the catchment evaluating alterna-
up with a strategy. OM were new in to change beef policy. Industry tives they may need to adopt in
their positions monitor farms future
Time Limited. Asked OM to put some Less restricted—mentor group as- Limited. Half day group meeting to
time in outside meetings with lim- pects part of another project. Re- identify criteria and alternatives.
ited success. Two group meetings. stricted to 2 h for individual Restricted to 2 h for individual
2 h individual meetings with OM. sessions, plus 1 h for evaluation sessions, including evaluation
Half-hour evaluation
Understanding Problem and technology not clearly Problem and alternatives readily Problem and alternatives readily
understood (not realised initially) understood understood
What was Identify the best strategy. OMs to Select a policy or reinforce current Identify and evaluate alternative
specified as take ownership of the problem and choice land uses. Understand what criteria
wanted? decision outcome (SM requirement) Explore decision and criteria. are important. Understand MCDA
Evaluate quantitative decision- process. Help out.
making process
Identify alter- None pre-defined. Report with None pre-defined. Brainstorm. 14 alternatives, most pre-defined by
natives economic evaluation done. Some Some identified by mentor group, Taupo researchers. Group sugges-
group discussion. Identified by OM. some by farmer. tions added. Worked in groups.
Identify Identified by OM. Value tree ap- Identified by farmers. Some mentor Defined at group meeting (14 cri-
criteria proach using revenue, cost and risk group suggestions. Brainstorming. teria). Worked in groups, then
branches. larger group.
Score Subjective scores Some subjective scores. Gross Some subjective scores. Gross
performance margin measure provided. margin and nitrogen leaching mea-
sures provided.
Weight criteria Various (swing weightings, trade- Swing weightings Swing weightings
offs, point allocation, ranking)
Process Survey. Interview OM. Observa- Interview or written response from Survey. Observations during pro-
Evaluation tions. managers. Observations. cess.

(Eden, 1995; Montibeller, 2007). However, this also regard to participants’ time availability and
meant the decision makers were in control— willingness to cooperate (in part dependent on
something that Eden and Ackermann (1998) note their perceived personal gain from the process).
as being important in dealing with clients. Hence, Barcus and Montibeller (2008) and Montibeller
effectively assisting the clients with their decision (2007) discuss methods for the evaluation of
making (i.e. their requirement) had to be as high a MCDA using action research. Our research was
priority, if not higher, than the research objective of action-led by necessity, but the incorporation of
evaluating MCDA. We also faced the same some more formal analyses was possible allowing
limitations as those planning and facilitating some general theory building. The participation of
MCDA for commercial clients, particularly with the researchers in the process and the inclusion of

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda
42 A. E. DOOLEY ET AL.

method selection by the researchers as part of the meeting (operational and senior management) to
case studies contributed to this being action evaluate technology use across all areas and come
research (Massey and Hurley, 1999; Montibeller up with a combined overall strategy. As the
2007). agribusiness case study was conducted with a
Evaluation methods were largely dictated by commercial company, the decision itself cannot be
the clients’ time availability and differed between discussed in any detail because of confidentiality
the case studies (Table I). Note that the number of issues. However, learnings from the application of
people was too small for the quantitative data MCDA can be discussed.
(surveys) to be statistically significant. The case
studies are described next. 2.2. Beef case study (FB case study, i.e. farm
business)
2.1. Agribusiness case study (AB case study) Meat and Wool New Zealand use farmer levies to
A large meat processing company wanted to fund research, extension and marketing activities
identify the strategy to best utilize a new objective in the sheep and beef farming sectors. This
carcass measurement technology developed in con- includes funding monitor farms which are com-
junction with AgResearch. Operations managers for mercial farms whose owners have agreed to work
procurement, processing, and marketing were dele- with researchers to investigate some aspect of their
gated the task of selecting a strategy by senior farming system. These farms act as showpieces and
management. They were to each consider the same focal points for discussion on what can be done or
decision, but focus on their own area, although how systems can be improved, predominantly for
the final strategy selected would be a senior local farmers, although articles are also written in
management decision. The operations managers the NZ farming press and included on their
were reluctant participants in that they wanted website and publications.
nothing out of the process other than a decision to AgResearch scientists were working with eight
present to senior management. Senior management monitor farms throughout New Zealand selecting
wanted to include operations managers in the the most suitable beef cattle policies. Each farm
process to get them to take some ownership and had a mentor group of 10–15 local farmers and
responsibility for the decisions regarding implemen- agribusiness people who provided ideas and
tation of the technology, and ensure their commit- opinions, e.g. alternatives. The ultimate decision
ment to the decision. on what to implement remained with the farm
The operations managers had been at owners. For this research, two of these monitor
the company for varying periods of time farms (one beef cow policy decision, one fattening
(6 months–10 years), and had different base cattle policy decision) agreed to incorporate an
levels of knowledge and skills. They worked MCDA component to assist them in making their
together effectively and were accustomed to decision.
making frequent decisions about operations, but MCDA integrated well into this project.
had not been involved previously in strategic Discussion of MCDA, and identification of
decision making. The technology was new and alternatives and some initial criteria were
unfamiliar to most of the managers, although incorporated into a mentor group meeting. The
some had been involved with the project to researchers then worked with the two individual
develop the technology. Despite being provided case study farmers (one a farm manager, the other
with a report on the technology use and the a couple farming in partnership) to quantitatively
economic benefits of different strategies, it became evaluate the alternatives. While the mentor group
apparent during the process that they had not and researchers had made some suggestions, the
understood the technology. owners identified and used criteria (including
There was an initial group meeting with subjective criteria) that were important to them.
operational and senior managers to discuss Most of the scoring against the criteria was also
different strategies and criteria. The researchers subjective and reflected their opinions. This
then worked individually with the operations MCDA approach was liked by the mentor
managers to evaluate the strategies for use in groups and participating farmers, and was
their area. They identified the criteria important to subsequently extended to seven of the eight
them in their area. Scoring was subjective. monitor farms (Smeaton and Dooley, 2006;
Individual meetings were followed by a group Meat, 2007).

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda
APPLICATION OF MULTIPLE CRITERIA DECISION ANALYSIS 43

2.3. Taupo case study (AE case study, i.e. assessed by the participants. A group meeting of
agricultural–environmental) stakeholders was initially held to identify criteria,
Lake Taupo is New Zealand’s largest lake. There followed by individual meetings with participants to
is concern that nitrogen contamination of Lake score a set of farm systems against the criteria and
Taupo is leading to slow deterioration in the lake’s then weight the criteria to identify the best systems.
water quality. One of the main, manageable There were 14 alternatives and 14 criteria. An
sources of nitrogen impacts to the lake is nitrogen attempt was made to develop a constructed scale
leaching from agricultural land. Consequently, for some criteria but participants found this difficult
farmers in the Lake Taupo catchment are faced to agree on and expressed a preference for subjective
with specific legislation from the local council scoring. Time available was limited, so participants
(Environment Waikato) to stop any further could reject outright an alternative stating a reason,
increase in nitrogen leaching, with nitrogen leach- or ignore criteria (zero weight) that were not
ing outputs for farms in the catchment around the important to them in order to reduce the number
lake to be capped. The farms affected are of assessments required. Subjective weighting and
predominantly sheep and beef farms. Some are scoring meant that direct comparisons between
privately owned owner-operator properties, while people in their assessments could not be made.
others are tribal land owned by Maori trusts and However, the alternative rankings and reasons they
incorporations (most of which cannot be sold). rated well or otherwise could be compared. This
AgResearch has been working with these study is discussed in depth in Dooley et al. (2005).
farmer groups to evaluate environmental impacts
(potential nitrogen leaching) from different
farming systems. However, if systems are to be 3. MCDA METHOD SELECTION AND
adopted they have to be acceptable to farmers. EVALUATION
Criteria such as profit, risk, and lifestyle are also
important to farmers in selecting between farm 3.1. MCDA method selection
enterprises. Therefore, it is important to In this paper we adopt the definition of MCDA as
understand what farmers consider when a process (Belton and Stewart, 2002), rather than
developing alternative systems or new just the aggregation procedure used. MCDA
technologies to reduce, or maintain nitrogen approaches differ in the way the information on
outputs. MCDA was used to identify and alternatives and their performance, and criteria
compare what some Lake Taupo stakeholders and their relative significance, are elicited, speci-
(including environmental council members and fied, and analysed (DTLR, 2001; Belton &
scientists) considered important in selecting Stewart, 2002). However, the generic process
between farming systems and technologies to described below, is similar for many approaches.
reduce nitrogen pollution in a capped nitrogen
environment. While this was not a ‘real’ decision in 1. Identifying the decision problem, context, and
the true sense of the word, the researchers and those involved.
farmers were working together on developing and 2. Identifying alternatives.
evaluating different technologies and systems, 3. Identifying criteria (what is important in mak-
which the farmers in this area will have to ing the decision).
consider adopting within the next few years to 4. Quantifying the importance of the criteria
comply with the capped nitrogen leaching (weights).
environment. This work continues. 5. Quantifying the performance of the alternatives
Comparisons were made between individuals on the criteria (measures or subjective scores).
and groups (privately owned, Maori Trusts, 6. Ranking the alternatives or identifying the best
researchers, council staff) in criteria, criteria alternative.
importance, scores, and farm system rankings.
A comparative group approach was taken (Belton & The most appropriate approach or methods
Pictet, 1997) allowing diversity between decision for a problem is dependent on decision-maker
makers in criteria, weights, and scores to be requirements, method requirements, and pro-
identified. A common set of agreed alternative blem characteristics (Tecle, 1992; Hanne, 1999).
farm systems and criteria were used, but weights, A comprehensive literature review of MCDA was
and most scores or measures were individually undertaken to identify the strengths, weaknesses,

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda
44 A. E. DOOLEY ET AL.

potential uses, and restrictions associated with It is recognized as being suitable for groups and
each of the predominant MCDA methods useful for facilitating learning (Belton and Stewart,
(Dooley, 2003, unpublished). This spanned the 2002). It is relatively quick, requiring fewer
MCDA, social sciences (including decision comparisons than similar methods such as AHP.
sciences), operations research and mathematics, Its popularity may partly be the result of many
and agriculture and environment literature. Key decisions having characteristics which lend
factors that needed to be considered in matching themselves to this method. The quantification
MCDA methods to the decision under and analysis stages of the process were similar
consideration were identified and incorporated for the case studies, particularly the FB and AE
into a descriptive framework to assist in selecting cases (Table I). An additive value function model
the most appropriate MCDA approach and was used to combine scores and weights. Weights
methods for a given problem. The most (in most cases) were derived using the swing
appropriate methods to use at different stages weighting method developed by von Winterfeldt
can become increasingly evident as the MCDA and Edwards (1986) in their ‘simple multi-attribute
process progresses, making this an on-going rating technique with swings’.
process. Logical Decisionss software was used for the
While all the three case studies were from the AB case study. A spreadsheet model was
agricultural sector, they had very different developed in Microsoft Excels for the FB and
combinations of problem types and decision- AE case study analyses. This included output
maker requirements, individual and group graphs to display results from the model sensitivity
decisions, decision makers with similar or analyses. The use of Microsoft Excels software
different objectives, and familiar versus new allowed these models to be left with the farmers
technologies. The case study attributes are for further exploration.
described in Table I. The farm systems scientists and social scientists
The following characteristics were similar for all involved in the projects already had a good
the case studies. relationship and a sense of rapport with the
farmers and business managers, which was an
 alternatives in all case studies were discrete; advantage in working with them to include an
 subjective judgements were required; MCDA component in their projects. They were
 there were limitations on time available; skilled in facilitation and working with farmer
 it is relatively easy to use and understand; groups, and assisted in working with groups and
 alternatives are ranked, identifying the best individuals to identify alternatives and criteria
alternative; (identifying alternatives was already a component
 it is transparent, which was an advantage of their own projects). These researchers used
particularly for the AB and AE case studies, techniques they were familiar with, such as
although this was not specified as a requirement brainstorming, breakout groups (AE), and
beforehand; ranking in this process.
 software was available (AB) or readily devel-
oped (FB and AE). 3.2. Evaluation of MCDA
An effective decision support tool or process
However, the three case studies differed in other delivers close to what it has been designed to do
respects, e.g. group versus individual decision (Eden, 1995). Decision makers evaluated the
making and interactions, understanding of the MCDA process used to assist with their decision
decision, willingness to be involved, decision- making against their expectations and require-
maker objectives, and alternative and criteria ments, and any other observations they had
identification (Table I). regarding the usefulness (or otherwise) of the
The multi-attribute value theory (MAVT) process. Key issues that need to be considered
approach described in Belton and Stewart (2002) when implementing MCDA were identified.
was used for all the three case studies. The Evaluation methods and questions asked were
similarities between the case studies, particularly similar to those used in other MCDA and decision
the top three factors, contributed to the selection support comparisons and evaluations (Evans and
of this method for all cases. MAVT is one of the Riha, 1989; Hobbs et al., 1992; Gunderson et al.,
more widely used MCDA approaches in practice. 1994; Zapatero et al., 1997; Qureshi et al., 1999;

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda
APPLICATION OF MULTIPLE CRITERIA DECISION ANALYSIS 45

Bell et al., 2001; Lai et al., 2002). Participants were Environment or context
asked for their opinions on: the suitability of • time limitations were stipulated in advance for all case studies, although
some were less restrictive than others (FB>AE>AB);
MCDA for the decision; the importance of some • the case studies were able to be incorporated within or alongside another
of the benefits of MCDA methods; how well the project (FB>AE>AB);
• the decision was an “individual” decision rather than a group decision
process met what they considered important in requiring less co-operation between decision makers (FB=AE>AB);
evaluating their decision and any other benefits or • greater flexibility was possible e.g. the decision maker was able to decide
on some aspects of the process to best suit their way of thinking
drawbacks they identified; their perceptions of (FB>AE=AB);
each stage of the process (e.g. effectiveness, • the alternatives were easy for the participants to arrive at (AE>FB>AB);
• the criteria were easy for the participants to decide on (AE>FB>AB);
understandability, timeliness, and ease of use);
and whether the decision arrived at was what was People
• participants were willing to be involved, approached the process in a
intuitively expected. positive manner and wanted to make a change (FB>AE>AB);
Evaluation data were collected using written • the participants were easy for the researcher-facilitators to relate to
(especially initially), making it easier to determine how best to
questionnaires (especially where time was limited), implement the process (FB=AE>AB);
questioning and observation of the participants • the decision makers were quantitative people making it easier for them to
relate to the MCDA approach (AB>=FB=AE);
during the process, and interviews (Table I). • decision makers were accustomed to making decisions similar to the one
Written questionnaires required about 15 min to they were evaluating (FB=AE>AB);
complete, and primarily consisted of questions Planning and implementation (may also relate to the context)
requiring answers on a Likert scale to enable • there was good understanding between those facilitating, which
contributed to a more confident and effective delivery overall
collection of the maximum amount of data in a (FB=AE>AB);
limited time. A few open-ended questions • The decision makers knew most of those involved in facilitating the
MCDA process (same for all);
requiring written replies were asked (e.g. least- • the decision required was understood by all those involved e.g. decision
liked aspect, best-liked aspect, what was wanted). makers, facilitators (FB=AE>AB);
Interviews of half an hour to one hour were • the technologies or alternatives were understood by all those involved
(FB=AE>AB);
conducted with some AB and FB decision • the process was understood by all those involved (FB=AE>AB);
participants who were prepared to commit time • the analysis process (and software) was interactive, and relatively
straightforward to understand and use to help facilitate the process (same
to do this. Numbers within the participant groups for all);
were small limiting the use of statistical • the outputs were designed to be easy for the decision makers to
understand and included graphics to display the results (same for all).
evaluation, but trends between and within the
groups will be reported. The participants were Figure 1. Differences between the case studies.
forthcoming in making spontaneous comments on
the process and its potential use, particularly
during and after the quantitative evaluation stage 4.2. Expectations and benefits
where the facilitator/researcher worked with one The required outcomes wanted by the decision
to two participants at a time. This added rich makers varied, and differed between, and within,
information to the results. Triangulation of case studies (Table I). These outcomes were
methods (survey, interviews, observation) and achieved in some cases, and overall most partici-
researchers assisted in comparing and validating pants were relatively satisfied with the results. The
data (Patton, 1990). AB case study decision had still not been finalized.
Some AE participants had hoped to learn more
about possible alternatives.
4. EVALUATION OF THE MCDA The benefits identified in this study have also
PROCESS been identified in previous MCDA literature. The
key benefits identified in these studies were
4.1. Case study contrasts and similarities learning and understanding, a structured process,
Differences in planning and implementation, en- and transparency. Participants identified benefits
vironment or circumstances, and people factors from MCDA other than those they initially
contributed to the way participants viewed the specified as being important (Figure 2). Many of
process. Some differences and similarities between the benefits were identified after the process, rather
the case studies which affected the success of the than prior to the process. While the first three
decision processes used are described in Figure 1. unspecified benefits (understanding, structured
The case study types are shown in brackets with the process) were objectives for some participants,
case where this situation was most advantageous they were identified by others as specified benefits.
(or least disadvantaged) presented on the left. Where a consensus group decision was required

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda
46 A. E. DOOLEY ET AL.

Unspecified benefits from working through the process which participants have the biggest impact on rankings, than in
identified included:
• a greater understanding of their own decision making e.g. what was
challenging the ranking itself. In some cases
important to them in decision making; decision makers in the FB case study accepted
• a better understanding of others’ perspectives; that the top-ranked alternative was not their
• a structured, quantitative means to work through a decision;.
• quantification which contributed to keeping the decision process intuitive preference, and even agreed after further
objective; exploration that the rankings and the analysis were
• a means of documenting the decision;
• an objective decision process which provides a more transparent means correct. This led to one FB participant re-
to explain the decision or stimulate discussion of alternatives (e.g. to
trustees, mentor group, senior managers);
evaluating his current decision-making process
• changing the way they thought about their decision making; (i.e. questioning whether it was too emotional
• validation that it was acceptable to consider non-profit criteria; and determining to be more objective in his
• a better understanding of where their area fits in, in the overall picture;
• getting together to discuss the decision, share ideas and come to an decision making going forward).
agreement; While learning per se is not generally seen to be
• a better understanding of the decision or alternatives, sometimes as a
result of identifying that more information was required and seeking the purpose of MCDA, it has been identified in the
that information. literature as a key benefit of MCDA (Belton and
Stewart 2002). Almost all the case study
Figure 2. Benefits participants identified from using the
participants indicated that they had benefited
MCDA process.
from some aspect(s) of learning, whether they
specified this as an objective or not (i.e. the
decision and technologies, criteria and their
(AB case study), ranking, understanding others’ importance, their own and others’ perspectives).
perspectives, understanding which criteria have the The impact on participants’ learning was also
biggest impact and giving participants a sense of reflected in the terminology they used to discuss
ownership were more important than they were for MCDA, i.e. words such as learn, explore, and
the more individual decisions (AE case study) understand were all used, often synonymously.
where participants were more interested in Hence, these case study results strongly support
exploring the decision. learning as the key benefit of an MCDA analysis.
MCDA is often regarded as a process or The participatory nature of the process was
quantitative tool to assist with decision making beneficial for both group and individual decision
by identifying the best alternative(s). Ranking making (although it should be recognized that
alternatives or identifying the best alternative even the ‘individual’ decision making often
was specified as an objective by the AB included more than one person). The greater
participants. This was less important in the FB understanding of the decision, and their own and
and AE case studies, particularly the AE case each others’ perspectives gained through working
where the decision was more futuristic. Despite together during the process was particularly
this, no-one appeared to perceive the MCDA important in the AB case studies where a group
process as necessarily being expected to identify consensus was required. However, even in the
the best alternative. They were not concerned if the other case studies, participants who worked in
alternatives ranked differently from their intuitive groups for some aspects commented that their
expectations, although their intuitive preference increased understanding of what was important to
was usually in their top two alternatives. Instead, others stimulated their own thinking. They
an unexpected ranking led them to question their enjoyed this aspect of the process.
thinking on the decision, and to explore further to A means of structuring the problem, allowing
ensure whether their inputs really reflected their participants to make more thorough decisions and
views. That is, is the model requisite? (Phillips identify what was important in their decision
1984). Their final decision was not necessarily the making, was regarded as desirable, and was
top-ranked alternative. This suggests that concern specified as a required outcome by some. The
over accuracy of MCDA results, which is often case study participants liked the structured
raised in the literature, is not as important as some approach, finding it helped them to think through
other factors in implementing MCDA, e.g. time and understand the decision, confirming what is
available, ease of use. recognized in the literature, i.e. that the structured
The participants in these case studies were MCDA process provides greater understanding
more interested in exploring the decision to (own and group), objectivity, and transparency
understanding which criteria (scores and weights) (Belton and Stewart, 2002). This structured process

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda
APPLICATION OF MULTIPLE CRITERIA DECISION ANALYSIS 47

contributed to identifying gaps in peoples’ MCDA no doubt contributed to some of the


knowledge. For example, it was realized that the difficulties experienced.
technology was not understood when working The decisions reported here were all strategic
through the process in the AB case study business decisions (agribusiness, farm business)
prompting further learning about the technology: where ‘time is money’. Time limitations were
it could be argued that this recognition and specified for all three case studies, and were a
learning about the technology was the greatest key factor in MCDA method selection and
benefit from MCDA in this case study. implementation. These time limitations affected
Most participants were accustomed to working what was able to be achieved. However, more
with numeric data and preferred structured or thorough, accurate but time-consuming methods
semi-structured decision-making processes. This are unlikely to have increased client satisfaction—
may have meant the structured, quantitative and may in fact have had the opposite effect. In
MCDA process was easier for them to use than contrast, reported MCDA case studies are often at
some people would have, and may have influenced the policy level, where processes are considered
their opinion of MCDA, which was largely more, with greater emphasis on ensuring the
positive. Hence, MCDA may be more useful involvement and understanding of a wide range
with similar types of people. However, many of participants. Time may be less of a limitation
people in the farming community would be relative to other requirements in these studies.
similar to our case study participants, suggesting Those working through the decision in the FB
MCDA could be applicable for agricultural and AE case studies agreed to commit time to this
business decisions. process. However, in the AB case study senior
Some decision makers observed that the managers committed operations managers to
objectivity of the structured process had the working through the process. This created some
potential to reduce the opportunity for the decision problems. The operation managers commented
to be swayed by some individuals. While most that senior managers should have recognized that
considered this to be important, three participants they needed more time available to work on the
commented that sometimes it is necessary for decision and stated this would have resulted in
an individual to step in and make the decision: them being more willing to participate and
while it is important in a group decision to have a achieved better results. This demonstrates that
decision-making process where all can contribute, company situations require a commitment
leadership is also required. The literature also (including time) to the decision and the process
supports this. at higher levels within the company, as well as at
The objective, quantified and structured MCDA the level the decision is being made. Schein (1999)
process provides a potentially transparent way to made similar observations and recommends not
explain the decision to others (Belton and Stewart, getting involved in facilitating a decision unless the
2002). This was also recognized in hindsight by the decision maker or their company is prepared to
case study participants who were accountable to make a commitment up front, preferably during a
others or were involved in group decisions (e.g. paid meeting to ensure they are serious.
couples; managers or consultants answerable to Participants in the AB case study had
senior management or boards of trustees, corporate difficulties in understanding the decision, and the
farm trustees). new and unfamiliar technology, demonstrating
that time needs to be allowed for early on in the
4.3. Challenges in implementing the MCDA process process, including during planning, to ensure there
Possibly the greatest learning from this study was is a common understanding (by facilitators and
in relation to the implementation of MCDA, planners as well as participants) of the decision
particularly with the AB case study which was and technologies. While some prior information
more difficult than the FB and AE case studies was available, not all participants had read or
(Figure 1). The decision context, people involved, understood this. This affected their ability to
the process used and its implementation all affect understand and assess its impact. These problems
how smooth a decision-making process will be. in understanding were exacerbated by lack of time
The main difficulties experienced in these case for explanations early on, and the assumption that
studies are related to time, understanding, and people already understood the technology. When
ownership. Our inexperience in implementing this lack of understanding was realized part way

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda
48 A. E. DOOLEY ET AL.

through the process, further information was The need for trust can affect method selection
required. Thus, time ‘saved’ early on in the and facilitation. This requirement for trust may be
process, did not save time in the long run! More more important with sensitive decisions (AE case
time is required where proposed alternatives or study) or agribusiness decisions (AB case study)
technologies are new. In contrast, the FB and AE where information may be less readily available,
case studies were more successful than the AB case e.g. internal company information, people issues
study. Factors contributing to this included: the within the company. Case study identification, an
decision process fitted more closely with on-going understanding of the problem context, and gaining
work; the process was explained more clearly; the decision makers’ trust was achieved by our
more information and time was available for the working with AgResearch staff who already had
researchers to learn about the context and get to established relationships with the company or
know those involved; and the participants were farmer decision makers. AgResearch facilitators
already familiar with many of the alternatives. probably had the advantages of both internal
Differences between people need to be facilitators (understanding of company and people
understood and allowed for when planning and involved) and external facilitators (neutrality)
implementing the process as these can contribute (Mayon-White 1990). While the methods used
to differences in their understanding. Even while had to suit all those involved, including the
working with a company decision such as in the AgResearch staff, the case studies would have
AB decision where those involved had similar taken considerably longer had these AgResearch
overall objectives, there were differences in their staff not been involved.
understanding, objectives and perspectives because
of their differing backgrounds, experiences, 4.4. Understanding and using the MCDA process
responsibility levels, and personalities. Participants found the qualitative aspects of the
The decision process and the need to be process (understanding the problem, identifying
accountable acted as a catalyst in the AB case criteria and alternatives) more difficult than the
study resulting in most managers having a greater quantification stages of the process. The problem,
sense of ownership of the decision at the end of the technology, and process were not clearly under-
process. While senior management in these case stood by the AB participants. However, the
studies wanted the operations managers to take process used contributed to identifying that there
some ownership and responsibility for the were misunderstandings about the technology.
consensus group decision, the operations These aspects will be discussed further later. The
managers were reluctant to be involved. They AB participants had to identify both the criteria
were busy, were not committed but had to and alternatives, and they disagreed or were
participate, were unaccustomed to this level of neutral as to whether this was easy to do. The
decision making, or saw no reason not to do things technology was new, and they had less time and
the old way (senior management made decisions at help available to assist with this than the other case
this level). This created some difficulties getting study participants. Managers who had put extra
them to actively participate in the process initially. time into this found it easier.
However, the later realization that they did not In contrast, most FB and AE participants
understand the technology sufficiently to make a found the information on the problem, process,
decision to present to their managers resulted in and alternatives (AE) to be adequate. FB decision
these operations managers taking initiative makers had the assistance of a mentor group to
between formal meetings to discuss the problem assist with alternative and criteria identification,
between themselves. In this respect the MCDA and more familiar technologies to consider. Even
process was a success: one of the earliest signs of so, it still took considerable time and input to
success in getting commited to a change occurs arrive at a set of alternatives and criteria, although
when external consultants find there has been one FB participant noted that it was still quicker
increased management of the decision by a group than if they had done this without mentor group
through informal debate and discussion between assistance.
convened meetings (Mayon-White, 1990). The majority of the time spent on the MCDA
However, it should be noted that any structured process was in understanding the decision and
and interactive decision-making process could identifying criteria and alternatives. The time
have had the same effect as MCDA. required for this was underestimated, particularly

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda
APPLICATION OF MULTIPLE CRITERIA DECISION ANALYSIS 49

in the AB case studies. In the FB and AE case lifestyle’. They noted that it would be necessary to
studies, these stages were better integrated into have some guidance the first time it was used,
other ongoing activities and the technologies were particularly in weighting criteria.
more familiar.
Once the process was explained, most
participants had little hesitation about allocating 5. THE POTENTIAL FOR MCDA IN FUTURE
weightings and scores and accepting the results. AGRICULTURAL DECISION MAKING
The most difficult aspect of the quantitative IN NEW ZEALAND
analysis identified was understanding the swing
weighting method. Once explained, most people The findings in this study suggest there is potential
agreed it was easy to use. Where people worked for MCDA to be used in agricultural decision
together on weighting and subjective scoring (e.g. making in New Zealand to help balance economic,
farming couple, two managers, manager and environmental, and social or lifestyle goals in
trustee) they largely agreed on weights and strategic decision making. Agricultural and envir-
scores, and did not appear to have any difficulty onmental organizations, including AgResearch,
coming to a compromise when there was initial are well aware of the increasing need to trade-off
disagreement. agricultural and environmental goals. There is on-
Questions were not raised about the methods going work looking at trade-offs, ways to bring
themselves, although there was considerable about change, likely farmer responses to initiatives,
discussion about the results and further analysis farmer decision making, and new alternatives,
in most cases, e.g. revision of the weights and which may need to meet a range of criteria to be
scores. It was interesting to observe that they spent adopted. Currently much of this work is at the
as much time reviewing and revising their inputs policy level and there is interest in MCDA methods
and assessing the impact these had as they had on to assist. There has also been other farmer-targeted
setting up the model. This was despite the fact that work as a result of this MCDA study.
this was optional and they had specified that there
were time constraints! Previous findings report that  The selection framework developed was used to
the visualization of results is important in assisting help determine the best MCDA method for
peoples’ learning (Belton and Stewart, 2002) and further work undertaken at AgResearch. This
can support sensitivity analysis. This was supported included an outranking approach for use in a
in this study where participants found the visual decision support tool to help farmers select
presentation of the analysis results (graphs) and the forages (Snow et al., 2005) and a compromise
interactive approach helpful, contributing to their programming approach for water allocation.
learning and interest in the results.  As previously mentioned, the scientist in charge
Most case study participants would be pre- of the beef monitor farm project used MCDA
pared to use the process again, with 5 out of for five other beef monitor farms because it
14 AE decision makers preferring it to their current integrated well with what they were doing, and
decision-making process (eight were neutral, one he was impressed with the process noting that
disagreed). Not surprisingly, participants who were farmers liked it too. This work has been
more willing to be involved were more positive. reported at a farmer conference (Smeaton and
Some of the AB participants (operations managers) Dooley 2006) and written up in a Meat and
who had no choice about participating, and AE Wool R&D brief (Meat, 2007, Figure 3).
participants who were only ‘doing this to help’
perceived the process as being less useful or would While the case study participants agreed (some
not want to use this process to assist in future reluctantly) to use a MCDA process for their
decision making. Participants considered the decision making as part of their involvement in
process could be useful for future decision making another project, it is unlikely they would have
for themselves and for other agricultural decision agreed to use this otherwise. Yet most found the
makers, with one participant commenting that he MCDA process to be more useful or interesting
thought the process was likely to be considered than they expected with some key benefits
useful for farm decision making by about 50% of (learning, increased understanding, a structured
farmers, i.e. those who were ‘strategic managers and transparent process) generally not recognized
looking at farming as a business rather than a until after the process had been used. Hence, the

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda
50 A. E. DOOLEY ET AL.

our results suggest that the structured process and


learning associated with MCDA is more important
than the rankings, and that the decision makers
were not overly concerned if their preferred
alternative was not top.
The quantitative nature of MCDA results
means the outcome of the decision-making
process is explicit. This may be of concern to
people who feel there is potential for results to be
misinterpreted and misused by others. In New
Zealand, there is a perception by some that
farmers do not care for their environment.
Consequently, farmers around Lake Taupo were
sensitive to the possibility that their results could
be misinterpreted to reinforce the views of these
people. These types of concerns need to be
considered in sharing the results of an analysis
with people other than the participants. It also
highlights the importance of making the following
factors clear in presenting and discussing results.

 Some criteria are a requirement rather than


being traded-off. For example, in the Taupo AE
study, the weighting on environment was low
for some farmers because a minimum level of
environmental performance was expected be-
Figure 3. Meat and Wool R&D brief reporting the use fore an alternative was even considered, rather
of MCDA to help farmers make beef policy decisions.
than trading environmental criteria off against
other criteria. Above this acceptable level,
challenge is to identify how to best promote or improvements in environmental performance
implement MCDA methods to encourage people were small and not high priority, with some
to consider using them when applicable. Given the participants noting that they already contribu-
time commitment required and the need to step ted to improving environmental impacts. How-
outside their comfort zone, it is likely to be ever, this did not mean that the environment
difficult to get people to consider using these per se was of low priority.
methods unprompted. Reporting of MCDA  There can be differences between peoples’
applications and associated benefits in industry weights because of their different perceptions
forums and publications could alert people to the or alternative sets, e.g. the Taupo AE case
existence of these structured, quantitative methods study. Care needs to be taken in comparing
to help assist decision making. MCDA results between people, particularly
Once people have participated in MCDA and where evaluations are subjective rather than
recognize the benefits they may be more likely to measured. Ranges between the best and worst
use this again. However, the resources (tools, possible performing alternatives on the criteria
skills, expertise) need to be available. Farm and will affect peoples’ weights and scores. Simi-
business consultants may be well placed to offer larly, results only apply to the set of alternatives
MCDA as part of their toolkit, e.g. to assist clients being evaluated, or alternatives that perform in
with exploring alternatives and helping with the the same range as these systems: another
decision once alternatives have been identified and decision or set of alternatives may well require
some measures calculated. A concern with this different criteria and weightings.
approach is that lack of expertise may result in
results that are less accurate than they could be. The MCDA approach used in this study could
However, it is uncertain how important this is, i.e. be useful to help identify systems or technologies

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda
APPLICATION OF MULTIPLE CRITERIA DECISION ANALYSIS 51

farmers might adopt, and factors influencing their problems were encountered in the initial stages of the
decision. Although the number of participants in decision process and could apply to any formal
the AE case study was small and results cannot decision process, particularly those involving groups.
necessarily be considered applicable to all farmers While these issues are not discussed in much of the
within the catchment, the wide variation in results MCDA literature, they have been identified as being
(Dooley et al., 2005) demonstrated the diversity in important in the wider decision making and
farmer perspectives suggesting a range of different consultancy literature. Personal preferences and the
farm systems will be needed. Our results also decision makers’ willingness to be involved
found that some of those involved in the accounted for some of the variation in their
development and selection of agricultural systems opinion of the decision process, with those who
and policy do not understand farmers’ perspectives were more willing to participate being more positive.
as well as they could. If policy people and Planning and implementation could have been
researchers want to facilitate change in farming improved particularly for the AB case studies. The
practices then it is essential that they understand time, understanding, and ownership problems
what is important to those who have to make the largely occurred in the initial stages of the process,
change. MCDA could contribute to this. and the opinion was expressed that these problems
were likely to have occurred with any formal or
group decision-making process. This concurs with
6. CONCLUSIONS Schein (1999), who notes that most mistakes in
consultation tend to occur at the beginning.
The range of criteria involved in agricultural Similarly, many of the benefits identified could
decisions and the benefits identified in these case have resulted from any formal, structured or semi-
studies suggest that MCDA can be helpful for structured decision-making process.
assisting with agricultural and agribusiness deci- Key points identified by this study, which need
sion making. Most case study participants believed to be considered in planning and implementing
MCDA would be useful for agricultural decision MCDA in future include the following:
making. The process was judged to be useful by
most participants but could have been better  Ensuring that decision makers or their company
implemented. Most would be willing to use it are prepared to invest time in a decision. Those
again suggesting that the MCDA decision process involved should be informed of the extent of
was appropriate: Eden (1995) noted that a decision their involvement and have adequate informa-
process that is liked by users and achieves what it tion before the decision process commences.
is supposed to demonstrates credibility and ade- Extra time needs to be allowed if the technology
quacy. Participants recognized many of the bene- is unfamiliar.
fits only after being involved in the process. They  Understanding the problem context before the
identified learning about the decision and technol- decision process is planned. This may be more
ogies; what was important to them; and others’ difficult with company agribusiness decisions
perspectives as the key benefits. The structured where access to information may be more
decision process provided an objective means to limited for facilitators or analysts.
work through the problem, and a potentially  Understanding as much as possible about the
transparent way for them to explain their decision decision makers (attitudes, experience, skills,
to others. The MCDA process contributed to AB reason for participating) in the planning phase.
managers being more committed to the decision.  Ensuring that the problem, decision process,
The key problems experienced in these case technologies, and any pre-defined decision alter-
studies included: insufficient time to complete the natives are understood by all decision makers,
decision process properly; difficulty in understanding and to some extent by facilitators and analysts at
the problem, decision process, and technology (AB); the start. Do not assume these are understood.
and decision makers unwilling to take responsibility  Recognizing the importance of establishing the
for the decision (AB) or not committed because they trust of the decision makers (and company if
felt they had little to gain from participating (AB, applicable), and working with contacts to ensure
AE). These problems were associated with this if need be. This may require trade-offs between
implementation and constraints on what was working with those trusted by the decision makers,
achievable, and were inter-related. Most of these methods used and facilitator skills.

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda
52 A. E. DOOLEY ET AL.

 Appreciating that considerable time and com- and business people who participated in the case
mitment may be required to identify criteria and studies, and without whom the research could
alternatives, especially where these are not pre- not have been undertaken, is very much appre-
defined or when dealing with unfamiliar tech- ciated. This post-doctoral research was funded by
nologies. AgMardt (Agricultural and Marketing Research
 Being flexible in implementation and prepared and Development Trust) Grant number: 20392.
to revise the plan if required or if it is likely to be
advantageous. Some factors may not be under-
stood or may not arise until the process is
underway. REFERENCES
 Realising that not everyone will necessarily be
happy with the decision process and possibly Barcus A, Montibeller G. 2008. Supporting the alloca-
the outcome because of differences in person- tion of software development work in distributed
ality, background, objectives, and expectations. teams with multi-criteria decision analysis. Omega 36:
464–475.
Research by Barcus and Montibeller (2008) also Bell ML, Hobbs BJ, Elliott EM, Ellis H, Robinson Z.
2001. An evaluation of multi-criteria methods in
identified some of the above: problem structuring
integrated assessment of climate policy. Journal of
was critical; the need for commitment and trust; Multi-Criteria Decision Analysis 10: 229–256.
time required and the need to commit to this for best Beinat E. 2001. Multi-criteria analysis for environmental
results; learning through interaction with the model. management. Journal of Multi-Criteria Decision
MCDA has the potential to be useful for future Analysis 10: 51.
agricultural decision making. While MCDA is Belton V. 2001. The need for interaction and integra-
often seen as a process or tool to help identify the tion. Journal of Multi-Criteria Decision Analysis 10:
best decision, it has the potential to do far more 127–128.
than this. While there has been considerable Belton V, Pictet J. 1997. A framework for group
research on MCDA methods, it is unclear how decision using a MCDA model: sharing, aggregating
or comparing individual information. Revue des
to best promote or implement MCDA methods to Systemes de Decision 6(3): 283–303.
encourage people to use them. Benefits such as Belton V, Stewart T. 2002. Multiple Criteria Decision
learning need to be promoted in advocating the Analysis: An Integrated Approach. Kluwer Academic
use of MCDA. Promotion of MCDA may best be Publishers: Boston.
achieved as part of other extension or consultation Dooley AE, Smeaton D, Ledgard SF. 2005. Identifica-
activities by someone who has used the process. tion of Important Criteria in Farm Systems Decisions
The time and commitment required may deter Around Lake Taupo. New Zealand Agricultural and
people from considering a structured decision- Resource Economics Society, AERU No. 152: 27–35.
making process such as MCDA. Simpler MCDA DTLR. 2001. Multi-criteria Analysis: A Manual.
approaches are likely to be as effective in achieving Department for transport, local government and the
region. Available at: http://www.dtlr.gov.uk/about/
many of the benefits (e.g. learning, stimulating multicritreria [Accessed on 12 August 2002].
thinking, objectivity) as more complex ones, and Eden C. 1995. On evaluating the performance of ‘wide-
are more likely to be acceptable to decision makers band’ GDSS’s. European Journal of Operational
given time constraints. The opportunity exists for Research 81: 302–311.
MCDA research in these areas. Eden C, Ackermann F. 1998. Making Strategy: The
Journey of Strategic Management. Sage Publications:
London, Great Britain.
Evans GE, Riha JR. 1989. Assessing DSS effectiveness
ACKNOWLEDGEMENTS using evaluation research methods. Information and
Management 16: 197–206.
The AgMardt post-doctoral fellowship which French S. 1998. Decision making not decision theory.
Journal of Multi-Criteria Decision Analysis 7: 303.
allowed this research to be undertaken is gratefully French S. 2000. It’s good to talky. Journal of Multi-
acknowledged. Thanks to AgResearch for hosting Criteria Decision Analysis 9: 125–126.
the fellowship. Thank you to Rex Webby, Ian Gunderson DE, Davis DL, Davis DF. 1994. Can DSS
Knowles, Terry Parminter, Warwick Waters, and technology improve group decision performance for
Alan McDermott of AgResearch who helped with end users? An experimental study. Journal of End User
the case studies. The contribution of the farmers Computing 72: 3–10.

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda
APPLICATION OF MULTIPLE CRITERIA DECISION ANALYSIS 53

Hanne T. 1999. Meta decision problems in multiple RAC. 1992. Multi-criteria analysis as a resource
criteria decision making. In Multicriteria Decision assessment tool. RAC Research Paper No. 6, Resource
Making: Advances in MCDM Models, Algorithms, Assessment Commission, March 1992.
Theory, and Applications, Gal T, Stewart T, Hanne T Roy B. 1999. Decision-aiding today. In Multicriteria
(eds). Kluwer Academic Press: Boston, USA; 6.1–6.25. Decision Making: Advances in MCDA Models, Algo-
Hobbs BF, Chankong V, Hamadeh W, Stakhiv EZ. rithms, Theory, and Applications, Gal T, Stewart TJ,
1992. Does choice of multicriteria method matter? An Hanne T (eds). Kluwer Academic Press: Boston;
experiment in water resources planning. Water 1.1–1.35.
Resources Research 28(7): 1767–1779. Schein EH. 1999. Process Consultation Revisited: Build-
Lai VS, Wong BK, Cheung W. 2002. Group decision ing the Helping Relationship. Addison-Wesley: Mas-
making in a multiple criteria environment: a case sachusetts, USA.
using the AHP in software selection. European Journal Shadbolt N, Martin S. 2005. Farm Management in
of Operational Research 137: 134–144. New Zealand. Oxford University Press.
Massey C, Hurley E. 1999. Action research as a Smeaton DC, Dooley AE. 2006. Adoption of new
mechanism for client-driven development. Proceed- technologies or management systems on sheep and
ings of the 43rd Australian Agriculture and Resource beef farms. Proceedings of the New Zealand Grassland
Economics Society Conference and the 6th New Association 68: 229–235.
Zealand Agricultural and Resource Economics Society Snow VO, Knowles IM, Fraser TJ, McIvor S, Lovatt SJ,
(CD-ROM), Christchurch. Dooley AE, Finlayson JD. 2005. The M&WNZ
Mayon-White B. 1990. Problem solving in small groups: ForageMasters: Decision support for forage selec-
team members as agents of change. In Tackling tion. In MODSIM 2005 International Congress on
Strategic Problems: the Role of Group Decision Modelling and Simulation, Zerger A, Argent RM
Support, Eden C, Radford J (eds). Sage Publications: (eds). Modelling and Simulation Society of Australia
London, Great Britain; 78–89. and New Zealand, December 2005. University of
Meat NZ. 2007. Farmer Decision Making—A Tool to Aid Melbourne, 259–265. ISBN: 0-9758400-2-9. Available
the process. Available at: http://www.meatnz.co.nz/main. at: http://www.mssanz.org.au/modsim05/proceedings/
cfm?id 5 44&spid 5 96 [Accessed on 1 January 2008]. papers/snow.pdf. [Accessed on 23 June 2006]
Montibeller G. 2007. Action-researching MCDA inter- (Modelling and Simulation Society of Australia and
ventions. In Key-note Papers, 49th British Operational New Zealand).
Research Conference, Shaw D (ed.). The OR Society: Tecle A. 1992. Selecting a multicriterion decision
Melbourne, Australia. making technique for watershed resources manage-
Patton MQ. 1990. Qualitative Evaluation and Research ment. Water Resources Bulletin 28(1): 129–140.
Methods (2nd edn). Sage Publications: Newbury Park, von Winterfeldt D, Edwards W. 1986. Decision Analysis
California, USA. and Behavioural Research. Cambridge University
Phillips LD. 1984. A theory of requisite decision models. Press: Cambridge.
Acta Psychologica 56: 29–48. Zapatero EG, Smith CH, Weistroffer HR. 1997.
Qureshi ME, Harrison SR, Wegener MK. 1999. Evaluating multiple-attribute decision support sys-
Validation of multicriteria analysis models. Agricul- tems. Journal of Multi-Criteria Decision Analysis
tural Systems 62: 105–116. 6: 201–214.

Copyright r 2010 John Wiley & Sons, Ltd. J. Multi-Crit. Decis. Anal. 16: 39–53 (2009)
DOI: 10.1002/mcda

You might also like