The current issue and full text archive of this journal is available on Emerald Insight at:

https://www.emerald.com/insight/1741-0398.htm

Cognitive biases in MCDM Cognitive

biases in
methods: an embedded filter MCDM
methods
proposal through sustainable
supplier selection problem
Nurcan Deniz Received 16 September 2019
Revised 16 January 2020
Department of Business Administration, Eskisehir Osmangazi Universitesi, 7 April 2020
Eskisehir, Turkey Accepted 13 April 2020

Abstract
Purpose – Expert evaluation is the backbone of the multi-criteria decision-making (MCDM) techniques. The
experts make pairwise comparisons between criteria or alternatives in this evaluation. The mainstream
research focus on the ambiguity in this process and use fuzzy logic. On the other hand, cognitive biases are the
other but scarcely studied challenges to make accurate decisions. The purpose of this paper is to propose pilot
filters – as a debiasing strategy – embedded in the MCDM techniques to reduce the effects of framing effect, loss
aversion and status quo-type cognitive biases. The applicability of the proposed methodology is shown with
analytic hierarchy process-based Technique for Order-Preference by Similarity to Ideal Solution method
through a sustainable supplier selection problem.
Design/methodology/approach – The first filter’s aim is to reduce framing bias with restructuring the
questions. To manipulate the weights of criteria according to the degree of expected status quo and loss aversion
biases is the second filter’s aim. The second filter is implemented to a sustainable supplier selection problem.
Findings – The comparison of the results of biased and debiased ranking indicates that the best and worst
suppliers did not change, but the ranking of suppliers changed. As a result, it is shown that, to obtain more
accurate results, employing debiasing strategies is beneficial.
Originality/value – To the best of the author’s knowledge, this approach is a novel way to cope with the
cognitive biases. Applying this methodology easily to other MCDM techniques will help the decision makers to
take more accurate decisions.
Keywords Cognitive biases, Behavioral operations management, Heuristics, Debiasing, Decision-making,
Supplier selection, Sustainability
Paper type Research paper

1. Introduction
The quality of choice is a vital issue in which decision makers choose an alternative from a set
of options in multiple-criteria decision-making (MCDM) research. However the decision
maker’s relative preferences between pairwise alternatives or on elicited tradeoffs between
the criteria determine the result (Korhonen et al., 2018). As a real-life landfill location selection
problem, if the expert is a protector of wildlife, the expert cannot assign precise importance
degrees to some criteria (Ma et al., 2017). Contrary to the classic economic models which
assume that individuals act to maximize self-interest according to the “economic rationality,”
Simon’s studies proved that people make decisions that are inconsistent, that are not Pareto
efficient or that are based on normatively irrelevant factors. Therefore, it is advised to focus
on perceptual, psychological and cognitive factors that cause deviations from the “rational
man.” A human’s inability to evaluate decision alternatives simultaneously rather than
sequentially, tendency to select an alternative that is good enough rather than the optimal one
and use of simplifying rules or heuristics to reduce the cognitive demand of decision-making
are defined as decision maker’s weaknesses and limitations in the “bounded rationality”
Journal of Enterprise Information
Management
This study is presented in MCDM 2019 “25th International Conference on Multiple Criteria Decision © Emerald Publishing Limited

Making” which was held in _lstanbul, Turkey in June 16-21.

1741-0398
DOI 10.1108/JEIM-09-2019-0285
JEIM concept of Simon (Bazerman and Tenbrunsel, 1998). The source of the biases is the decisional
shortcuts known as “heuristics” (Braverman and Blumenthal-Barby, 2012). Tversky and
Kahneman showed that people rely on a limited number of heuristic principles, which reduce
the complex tasks to simpler judgmental operations. By the way, they are quite useful; these
heuristics lead to severe and systematic errors. Unfortunately, without being aware of them,
individuals frequently rely on these simplifying strategies. In terms of pros and cons, time-
saving and outweighing any potential loss resulting from a full-search strategy are the
benefits of heuristics that cannot be ignored (Bazerman and Tenbrunsel, 1998). But, they are
also guilty of causing cognitive biases that affect our decision-making. Cognitive biases are
important not only individual, but also organizational decisions. In the managerial context, as
a result of suboptimal or inefficient decisions depend on biases, the organizational value can
be undermined (Borrero and Henao, 2017).
As Korhonen and Wallenius (1996) state, integrating the results of the behavioral decision
theory into the design and development of interactive multiple-criteria methods is certainly
an area that has been overlooked in the operations research literature. The psychological
aspects of decision-making in operations research (OR) contexts are recently discussed in
behavioral OR (Korhonen et al., 2018).
The backbone of the MCDM techniques is the expert evaluation. The experts make
pairwise comparisons between criteria or alternatives. The mainstream research focus on the
ambiguity in this process. Fuzzy logic is the commonly used approach to overcome this
problem. On the other hand, cognitive biases is the other but scarcely studied challenge to
make accurate decisions. The first three heuristics described in the prospect theory (Tversky
and Kahneman, 1974) are representativeness, availability of instances and adjustment from
an anchor. Gains and losses are evaluated differently due to the shape of the value function,
which is concave for gains and convex for losses according to the prospect theory. Tversky
and Kahneman also demonstrated that choices can be influenced by the framing (i.e. wording)
of decision alternatives. Kahneman (2013) collected these information under “Thinking Fast
and Slow” book in which two decision-making processes – System 1 and System 2 – are
explained. This process is also known as the dual process theory (DPT) of decision-making
(Croskerry et al., 2013a, b). System 1 is rapid, instinctual and emotional. It is often on impulse
and is part of the fight or flight response. By the way, System 2 is slower, deliberate and
logical. Therefore, System 1 is presented as intuitive heuristic, and System 2 is presented as
deliberate analytic. Decision-making is a result of the interplay of these two systems (Osmont
et al., 2015). Although biases can occur in both types of systems, most biases are associated
with heuristics and typically are System 1 (Croskerry et al., 2013a, b).
According to the literature on cognitive biases related with MCDM, it is found that the
anchoring effect in interactive MCDM is studied by Buchanan and Corner (1997). However,
after 20 years, Borrero and Henao (2017) stated that the relationship between cognitive biases
and these decision-making tools has not been amply investigated. The main contribution of
this study is to propose embedded filters to avoid bias effects in MCDM. To the best of the
author’s knowledge, this approach is the first in the literature.
This paper advances the state-of-art knowledge in the field of MCDM in the following
aspects: (1) this approach can deal with the effects of cognitive limitations by introducing two
pilot filters; (2) these two filters help to reduce the cognitive limitations’ effects to make
decision more rationally; (3) more specifically, this approach can incorporate the framing
effect, loss aversion and status quo bias-type cognitive biases; (4) the first filter makes the
comparisons more balanced; (5) the second filter works according to the weight manipulation;
and (6) the approach is very flexible and simple, so it can also be applied to other MCDM
problems.
The remainder of this article is organized as follows. In Section 2, the literature is reviewed.
Section 3 introduces cognitive biases and debiasing methods. The proposed debiasing
methodology is presented in Section 4. In Section 5, the methodology is applied to a sustainable Cognitive
supplier selection problem. Lastly, the conclusions and future directions are given in Section 6. biases in
2. Literature review MCDM
This study is related with a variety of literature. As a consequence, this section is segregated methods
according to the issues. There is a subsection regarding sustainable supplier selection
problem, especially solved with MCDM methods. Cognitive biases and debiasing strategies to
reduce the effect these biases generally studied in psychology area. In this review, biases
related to studies in the decision-making area are investigated. The intersection of cognitive
biases and MCDM methods is analyzed in another subsection. There are some studies related
with more than one subsection, but they stated in only one to prevent duplication. Lastly, the
research gap is expressed in the last subsection.

2.1 Sustainable supplier selection problem and multi-criteria decision-making methods

Zimmer et al. (2016) analyzed and reviewed 143 peer-reviewed publications between 1997 and
2014 on sustainable supplier management (SSM). The results of this study showed the
predominance of the analytic hierarchy process, analytic network process and fuzzy-based
approaches. Also, the studies focused on the final evaluation and selection process step. Lastly,
it was found that social and quantitative metrics are neglected. Ma et al. (2017) extended the
analytic hierarchy process (AHP) method to cope with cognitive limitations in the MCDM
problem. Qian et al. (2019) developed a novel BOCR-uRTODIM solution method for multi-
attribute decision-making considering the buyer’s loss-averse behavior with incomplete
information. They combined an MCDM method (TODIM) with the prospect theory and used
their proposed methodology in clean energy devise procurement. Arabsheybani et al. (2018)
developed a novel integrated approach, including fuzzy MOORA and failure mode and effects
analysis (FMEA). They used the MCDM method called fuzzy MOORA to determine the
supplier’s score and FMEA to calculate the supplier’s risk. Stevic et al. (2020) developed a new
MCDM method called the measurement of alternatives and ranking according to COmpromise
solution (MARCOS) method for a sustainable supplier selection in the healthcare industry.

2.2 Biases and debiasing methods

Soman and Liu (2011) studied corrective strategies for the illusion of delayed incentives. After
the experiments, they showed that this bias of differential discounting can be corrected by
asking consumers to perform effort at the time of decision-making. Croskerry et al. (2013a)
reviewed the origins of biases in a special context of clinical decision-making. They also
outlined a general schema of how cognitive change occurs and examined the major strategies
for cognitive and affective (Croskerry et al., 2013b). Frydman and Rangel (2014) used a stock
trading laboratory experiment to analyze the possibility of reducing subjects’ tendency to
exhibit a disposition effect. To debias the disposition effect, information about a stock’s
purchase price is displayed. Cheng et al. (2014) examined the debiasing effect of “elaboration”
and “consider the opposite” to eliminate the framing effect on internet consumers’ decisions.
After the experimental results, they found that both strategies are useful, but “consider the
opposite” is more effective than “elaboration.” Morewedge et al. (2015) showed that debiasing
effects were domain general, and bias reduction with a single training intervention occurred
across problems in different contexts. Montibeller and Winterfeldt (2015) reviewed the
existing debiasing techniques to overcome the motivational and cognitive biases. As a result
of this study, a guidance is provided. Ma et al. (2017) proposed a novel approach that can
handle the cognitive limitations in MCDM problems by solving an estate investment problem.
They gave a measure to consider the effect of cognitive limitations and presented two
methods to reduce the influence of cognitive limitations. To reduce the confirmation bias via a
one-shot training intervention was the aim of Sellier et al.’s study (2019). The results of
JEIM experiments showed that debiasing-training effects can improve decision-making. Lee (2019)
explored the priming of a target attribute as an indirect debiasing method. After three
experiments, it was demonstrated that judgmental biases reduced in likelihood estimations:
ratio-bias and base-rate neglect when participants have enough cognitive resources.

2.3 Biases in multi-criteria decision-making methods

The literature about the intersection between (multi-criteria) decision-making and cognitive
biases is scarce. The first study accessed is performed by Buchanan and Corner (1997). The
aim of this study was to explore the anchoring effect in the interactive MCDM methods.
Almashat et al. (2008) focused on the effects of a debiasing procedure on three different
medical decisions involving cancer treatments. Morton and Fasolo (2009) reviewed
researches about BDT relevant to multi-criteria decision Analysis (MCDA). They focused
particularly on the multi-attribute value theory (MAVT). Cheng and Wu (2010) investigated
the warning effect as a debiasing strategy on framing. They also wanted to discover the
moderating role of individual involvement in decision-making. Braverman and Blumenthal-
Barby (2012) proved that clinicians’ decisions are affected from loss aversion, numeracy
toward frequencies, omission, confirmation and availability-type cognitive biases. The
results showed that more than 10% healthcare providers recommended to continue an
ineffective treatment. Kim et al. (2014) studied about choice, attribute and goal-type framing
effects. This study was an extension of previous single-alternative and single-attribute
situations to multi-alternative and multi-attribute situations. According to the experimental
studies, it was shown that attribute and goal framing influence consumer decision-making.
Osmont et al. (2015) studied ambiguity aversion’s effect on the framing effect through a
financial decision-making problem. Walmsley and Gilbey (2016) investigated cognitive
biases in visual pilots’ weather-related decision-making. On the other hand, confirmation and
desirability of choice biases’ effects on a decision made with MCDA techniques are studied by
Borrero and Henao (2017). The results show that the effectiveness of MCDA might be affected
by individual differences and corresponding motivational biases. Haita-Falah (2017)
researched the effect of the sunk cost fallacy and found that people are more committed to
a previously chosen alternative, if made responsible for that decision at an earlier point in
time. Guo et al. (2017) studied the time pressure interaction with framing effects in risky
decision-making. Consistent with the dual-process hypothesis, framing effects increased
under time pressure. Pandey and Jessica (2018) also measured behavioral biases affecting real
estate investment decisions in India. They used representativeness, availability, anchoring,
overconfidence and the gambler’s fallacy to measure loss aversion, regret aversion and
mental accounting biases. Shu (2018) studied bias effect on hiring a personnel. Korhonen et al.
(2018) proved that contextual variables in an MCDM task influence choice quality.

2.4 Research gap

However, cognitive biases actually play a vital role in MCDM methods; there have been few
researches on how to formally model the cognitive limitations in MCDM problems. Some
researches (Soman and Liu, 2011; Croskerry et al., 2013b, Frydman and Rangel, 2014; Cheng
et al., 2014; Morewedge et al., 2015; Sellier et al., 2019 and Lee, 2019) only focus on
psychological investigations and experiments, as summarized in Section 2.2. Only few of
them give a formal treatment for MCDM problems (Ma et al., 2017). This paper’s approach
tries to close the gap with a new stream of debiasing strategy. Manipulating weights and
forecasting the expected bias are novel approaches to reduce the effects of biases.

3. Biases and debiasing methods

Cognitive bias is defined as “a tendency to make systematic errors in certain circumstances
based on cognitive factors rather than evidence” (Braverman and Blumenthal-Barby, 2012).
According to the Arkes’ taxonomy, biases are classified in three classes based on their Cognitive
psychological origin: strategy-based (SB) errors, association-based (AB) errors and biases in
psychophysically based (PB) errors (Montibeller and Winterfeldt, 2015).
Anchoring is a PB type bias that is defined as a well-documented psychological bias in
MCDM
decision-making. Anchoring is related with the starting point of a decision process often methods
greatly affects the final outcome. After the experiments, it was shown that anchoring bias
present in MCDM (Buchanan and Corner, 1997). These experiments show that anchoring bias
affects the final decision in estimation tasks, pricing decisions and also in negotiations.
Decision makers allocate similar weights to all objectives in another PB-type bias called
“equalizing bias.” Studies showed that elicitation of weights in value trees and elicitation of
probabilities in decision trees are affected by this bias. Gains and losses may lead to different
answers in the context of gain–loss bias (PB type).
As an AB-type error, overconfidence bias occurs when the decision makers provide
estimates for a given parameter. This estimate can be above from the actual performance
(overestimation) or the range of variation is too narrow (overprecision). This bias generally
affects the quantitative estimates (Montibeller and Winterfeldt, 2015). In elicitation of weights
in multiattribute utility and value measurement, proxy attributes receive larger weights than
the respective fundamental objectives in accordance with the proxy bias (PB errors). Weights
of objectives are not properly adjusted to changes in the range of attributes in accordance
with the range insensitivity bias (PB errors), and splitting bias (PB errors) occurs when the
objectives are grouped in a value tree that affects their weights (Montibeller and Winterfeldt,
2015). It has been shown that judgments and decisions can be influenced greatly by whether
the information is presented or framed (Cheng and Wu, 2010). Risk, task and participant
characteristics are the factors that contribute to the framing effects (Huang et al., 2015).
Sunk-cost fallacy means people are more committed to a previously chosen alternative if
made responsible for that decision at an earlier point in time (Haita-Falah, 2017). The category
size bias demonstrates that people infer the characteristics of the category with its individual
members, and categorization affects probability judgments (Perfecto et al., 2018). Ambiguity
aversion is one of the most striking decision biases. In the context of this bias, a systematic
tendency is presented by decision makers to avoid ambiguous options for which the level of
risk is unknown (Osmont et al., 2015). Status quo bias is another widely documented bias and
also known as reference point effect (Morton and Fasolo, 2009). Affect heuristic presents
when people consider first whether the object of assessment is “good” or “bad,” and then
evaluate the disaggregate criterion scores on the basis of this overall assessment (Morton and
Fasolo, 2009).
Morton and Fasolo (2009) express four phases of the MCDA process: structuring,
assessing values, weighting criteria and sensitivity analysis. Different biases play a role in
different phases. Decision makers’ ability to generate criteria with bottom-up, top-down and
outside-in approaches are discussed under the structuring phase. On the other hand, the
scope insensitivity and status quo biases affect the assessing values phase. The behavioral
properties of different weighting questions and splitting bias in weight assessment are the
issues discussed under the weighting criteria phase. Lastly, the fourth phase about
sensitivity analysis is omitted in this study. It is possible to represent the decision makers’
preferences with weighting, but weighting criteria is the most cognitively demanding part of
the MCDA process. Consequently, predictable biases exhibit after the direct judgments about
weights (Morton and Fasolo, 2009).
Attempts to eliminate, or at least reduce, biases are called “debiasing.” Motivational and
cognitive biases can distort analysis inputs and are difficult to correct. To overcome these
biases, decision and risk analysis are designed (Montibeller and Winterfeldt, 2015). These
aforementioned biases can be classified as difficult or easy to correct. The overconfidence
bias, anchoring and insufficient adjustment and the equalizing bias are some of the examples
JEIM of difficult-to-correct biases. The literature on debiasing strategies focused on cognitive
biases is scarce. It was shown that the overconfidence bias can be reduced with the choice of
an appropriate elicitation technique (Montibeller and Winterfeldt, 2015). On the other hand,
conjunction fallacy and neglect of base rates are the examples of easy-to-correct biases. The
debiasing strategies change according to this classification. To eliminate biases that are easy
to correct, logic and decomposition are the most common strategies. As an example
demonstrating the probability logic can be used to eliminate conjunction fallacy. The neglect
of base rates can be corrected by eliciting base rates and conditional probabilities separately.
SB-type errors are easy to correct with the use of analytical models. By the way, AB- and PB-
type errors are difficult to correct (Montibeller and Winterfeldt, 2015). Cheng and Wu (2010)
prompted the participants with a warning message about the possibility of decision bias as a
debiasing mechanism. Cheng and Wu (2010) found a relationship between participant
involvement and effect of warning on message framing. As a result, it was shown that the
prompt of weak warning message is not an effective debiasing mechanism for less involved
participants. Also, there are evidence about informing people about a particular bias is not
useful (Almashat et al., 2008). Swing weighting and tradeoff weighting are alternative
elicitations for direct weight judgments. In swing weighting, the differences in value between
the levels of a most and least preferred option on two given criteria (swings) are made salient,
and respondents are explicitly requested to consider the relative value of the swings.
Respondents are requested to find an upper level of some underlying attribute that would
equilibrate swings on two designated criteria in tradeoff weighting (Morton and Fasolo,
2009). Buchanan and Corner (1997) presented the role of structures on the anchoring bias.
Because heuristics and biases are often adaptive, it is shown that training could impair
judgment and decision-making (Sellier et al., 2019). Croskerry et al. (2013b) outlined three
groups of suggested interventions: educational strategies, workplace strategies and forcing
functions.
Some debiasing techniques dependent on the bias type are proposed in the literature. As
an example, to provide multiple and counteranchors and to use different experts who use
different anchors are the suggestions for anchoring bias. On the other hand, it is advised to
rank events or objectives first, then assign ratio weights and to elicit weights or probabilities
hierarchically to reduce equalizing bias. Clearly identifying the status quo (SQ), expressing
values as marginal changes from SQ for value functions and eliciting utilities for gains and
losses separately for utility functions are the advices to reduce gain–loss bias. To provide
probability training, to start with extreme estimates (low and high), to avoid central tendency
anchors, to use counterfactuals, to challenge extremes and to use fixed value instead of fixed
probability elicitations are the debiasing techniques to reduce overconfidence. To overcome
proxy bias, to avoid proxy attributes and to build models relating proxies and fundamental
objectives and to provide weights for fundamental objectives are suggested. To make
attribute ranges explicit and to use swing weighting procedures, to use tradeoff or pricing-out
procedures and to use multiple elicitation procedures and cross-checks are the debiasing
techniques used to cope with range insensitivity bias. To avoid splits with large probability
or weight ratios, to use hierarchical estimation of weights or probabilities and to use ratio
judgments instead of direct estimation or distribution of points are the strategies to avoid
splitting biases (Montibeller and Winterfeldt, 2015).
According to the research, the elicitation of weights is influenced by splitting, equalizing,
gain–loss, proxy, range insensitivity, desirability of options and affect influenced bias. It is
proposed to obtain objectives and attributes from multiple stakeholders to reduce the
splitting bias. Evidently, engineers provide detail about cost and performance, and
environmentalists provide detail about environmental objectives. To set up the lower and
upper anchors of each attribute in a way, to use ranking and ratio weighting methods coupled
with hierarchical weighting are the approaches used to reduce the equalizing bias
(Montibeller and Winterfeldt, 2015). Soman and Liu (2011) classify corrective strategies of Cognitive
biases as debiasing and rebiasing strategies. biases in
MCDM
4. The proposed methodology methods
Two embedded filters are proposed as a debiasing strategy in this study. To counteract the
accurate decisions about cognitive biases is the aim of these filters. Framing effect (especially
goal framing), gain–loss bias and SQ effect are the cognitive biases tried to be reduced in
this study.

4.1 Filter 1
According to the “prospect theory” (Kahneman and Tversky, 1979), the wording of decision
alternatives affects choices by changing the reference point. When the same decision problem
is framed in different ways, it is shown that the relative attractiveness of options varies
according to the framing effect (Cheng and Wu, 2010). Buchanan and Corner (1997) suggested
that the choice of the starting point is important and influential. Korhonen and Wallenius
(1996) stated if allowed, decision makers may wish to reconsider alternatives once discarded.
The first reason of this cyclic behavior is the subjects’ comparison of alternatives in a relative
sense, rather than absolute sense. One of the other reason is subjects’ mistakes due to the
complexity of the decision task. Korhonen and Wallenius (1996) also indicate that the most
preferred solution may depend on the starting point and/or the path leading to the most
preferred solution. It is proved that the final choice is affected by “path” or sequence in which
solutions or settlements are presented to decision maker. According to the suggestions of
Korhonen and Wallenius (1996), to pay attention to the path and “framing” a problem are very
important. Taking into consideration this information and evidence, the first filter is
proposed to restructure the questions. Einstein’s “If I were given an hour in which to do a
problem upon which my life depended, I would spend 40 min studying it, 15 min reviewing it
and 5 min solving it” quote shows that the difficult part of decision-making is often not about
the right answer, but rather asking the right question (Ryan et al., 2018). To implement this
idea, it is proposed to rank the criteria in an alphabetical order in the first step. After this, it is
proposed to start not always with the starting point, but to change the starting point
sequentially to make the comparison pairs more balanced. The implementation of this two-
step approach is explained in the next section in detail.

4.2 Filter 2
MCDM problem’s solution involves two aspects: the evaluation of each decision alternative
regarding each criterion and the importance degree of each criterion. These two aspects need
to be investigated to model an MCDM problem with cognitive limitations (Ma et al., 2017). As
a result, the second filter depends on manipulating the weights according to the degree of
expected cognitive biases. Weights are defined as “scaling constants that represent value
tradeoffs and aggregate the partial values.” It is displayed that many common mistakes are
made in defining weights and several elicitation protocols for eliciting weights in an
appropriate way (Montibeller and Winterfeldt, 2015). Korhonen and Wallenius (1996) made
experimental research on individuals’ cognitive models. As a result of this study, it was found
that the relationship between weights and importance of criteria cannot be explained
exhaustively. In addition, Kim et al. (2014) showed that the weight of the framed attribute
becomes more important in the positive goal framing condition than in the negative. As an
example, it is proved that women decisions affected more than men decisions from cognitive
bias. Also, from the point of view that deciding faster will increase the bias risk, the response
time will be measured and will be used to manipulate the weights (Guo et al., 2017). Expertise
JEIM level and SQ effects are also quantified to manipulate the weights. The objective of the criteria
whether minimizing or maximizing is the other important factor to detect the gain–loss bias.
There is evidence about the dominance of cost which is a minimizing criterion (Buchanan and
Corner, 1997). After the extensive investigations on framing effects, it is shown that losses
loom larger than gains. This means, when alternatives are framed as gains, people display
risk aversion. On the other hand, people are risk-seeking when objectively equivalent
alternatives are framed as losses (Miu and Crişan, 2011).

5. Application of the proposed methodology through sustainable supplier

selection problem
“Supplier selection” is one of the most critical problems in supply chain management. The
solution of this problem is a cognitive process and characterized by a high degree of
uncertainty. MCDM techniques are commonly used in the solution of supplier selection
problem. Because the decision maker is not a rational economic agent, combining behavioral
operations management and decision-making literature is useful. Procedural rationality on
decision-makers’ perceived uncertainty, risk propensity, risk perception and perceived
supplier control are some of the behavioral issues studied in this context (Deniz, 2018). As a
response to increasing public awareness of environmental protection, supplier selection
problem evolved to “green” supply chain management problem. The actual version of the
problem is “sustainable supplier selection problem,” which is the combination of economic,
green and also “social” dimension (Stevic et al., 2020).
Haeri and Rezaei (2019) presented a real-world case study about green supplier selection
problem to show the applicability and effectiveness of their proposed model. A novel weight
assignment model is proposed by combining the best–worst method and fuzzy gray
cognitive maps model. Ten criteria are determined to determine the best supplier in this
study. Five of them are selected to show the proposed methodology simply: quality, cost,
delivery speed, pollution control and green image. Quality, cost and delivery speed are the
criteria represent “economic” dimension of sustainability. On the other hand, pollution control
and green image are the criteria for green dimension. “Worker safety” criteria (Zimmer et al.,
2016) is also added as a new criterion about social dimension used to enlarge the green
supplier selection problem to sustainable supplier selection.

5.1 Filter 1 implementation

Instead of the criteria order in Haeri and Rezaei (2019) article, six criteria are ranked in
alphabetical order as proposed in Table 1.
In a classic comparison of criteria, the decision maker answers the “How important is first
rank criterion (quality) when it is compared with second rank criterion (cost)?” question. In the
context of six criteria (n), there is a need to make 15 (n (n-1)/2) comparisons to compare all
pairs. As it can be seen in Column 2 of Table 2, in general comparison, the process starts with

Criteria number Criteria order (Haeri and Razei, 2019) Criteria order (alphabetical order)

1 Quality Cost
2 Cost Delivery speed
3 Delivery speed Green image
Table 1. 4 Green Image Pollution control
Criteria chosen and 5 Pollution control Quality
their alphabetical order 6 Worker safety
Comparison number Comparison pairs (actual) Comparison pairs (proposed)
Cognitive
biases in
1 1–2 1–2 MCDM
2 1–3 2–3
3 1–4 3–4 methods
4 1–5 4–5
5 1–6 5–6
6 2–3 6–1
7 2–4 5–2
8 2–5 4–6
9 2–6 3–5
10 3–4 2–4
11 3–5 1–3
12 3–6 6–2
13 4–5 3–6
14 4–6 4–1 Table 2.
15 5–6 5–1 Comparison pairs

the first criterion, and it is fixed. All the other criteria are compared with the first one. To make
the comparison pairs more balanced and reduce the framing effect, it is proposed to make
comparisons like in Column 3 of Table 2. The first criterion in the pair is always changing
sequentially in these alternative pairs.
As a result of changing the comparison pairs, the first and the second rank of the pairs
become more balanced. Table 3 shows this effect quantitatively. For example, the first
criterion is placed five times in the first rank in actual comparison, and this number decreases
with the growing number of criterion. Also, the first criterion is never placed in the second
rank in actual comparison, and this number increases with the growing number of criterion.
In accordance with the framing effect, this can result bias in decision-making. By the way,
with the proposed pairs, each criterion is placed in both the first and second rank two and
three times, respectively. These results show the proposed filter’s objectivity.

5.2 Filter 2 implementation

To show the second filter’s effectiveness, a biased ranking is obtained with Technique for
Order-Preference by Similarity to Ideal Solution (TOPSIS) (Hwang and Yoon, 1981) using
initial weights determined with the AHP (Saaty, 2016). These methods are chosen according
to their wide usage (Zimmer et al., 2016). Because the main issue of this study is a debiasing
mechanism, there is not given any information about the AHP and TOPSIS process. In the
second step, weights are manipulated as introduced before, and a debiased ranking is

Actual Proposed
1st rank 2nd rank 1st rank 2nd rank

1 5 – 2 3
2 4 1 2 3
3 3 2 3 2
4 2 3 3 2
5 1 4 3 2 Table 3.
6 – 5 2 3 Effect of proposed
TOTAL 15 15 15 15 comparison pairs
JEIM obtained with the same MCDM techniques. The workflow diagram for Filter 2 is visualized in
Figure 1. The details are given below.
5.2.1 Biased ranking.
First step: determining the criteria weights with analytic hierarchy process
The weights of six criteria are determined with the AHP technique. As a result, 15 binary
comparisons are made by the author, and the weights of the six criteria are determined,
respectively, as 0.204, 0.323, 0.076, 0.179, 0.142 and 0.077.
Second step: obtaining supplier evaluations
Expert opinions are derived from the study of Haeri and Rezaei (2019) article in which five
experts are ranking five suppliers. Each expert’s evaluation expressed with linguistic terms
are transformed to numbers. For example “very good” statement transformed to 5, and
“good” statement transformed to 4. The other transformations are made as “medium” 5 3,
“poor” 5 2 and “very poor” 5 1. As an example, the second expert’s evaluations in terms of
numbers can be seen in Table 4.
Third step: generating group decision
The next step is getting the average value of each expert’s evaluation as a group decision.
This means all experts’ decisions are equally weighted. Because there are five experts, each

Biased Ranking Debiased Ranking

1. Determining criteria
1. Manipulating criteria 2. Determining experts’
weights (AHP)
weights (loss-aversion) expected bias degree

2. Obtaining supplier 3. Obtaining supplier

evaluations evaluations
(Status quo bias)

3. Generating group 4. Generating group

decision (equally decision (biased
weighted) weighted)

Figure 1. 4. Ranking Suppliers

Workflow diagram for 5. Ranking Suppliers
(TOPSIS)
Filter 2 (TOPSIS)

Second expert S1 S2 S3 S4 S5

1 – Quality 3 4 2 5 1
2 – Cost 2 2 2 4 3
3 – Delivery speed 3 4 2 3 3
Table 4. 4 – Green image 3 2 5 3 3
Second expert’s 5 – Pollution control 3 3 4 2 3
evaluations 6 – Worker safety 2 3 3 2 3
expert’s weight is determined as 0.2 (1/5). As a result, Table 5 is generated as the input for Cognitive
TOPSIS. biases in
Fourth step: ranking suppliers with Technique for Order-Preference by Similarity to Ideal MCDM
Solution methods
TOPSIS is a common MCDM technique used for ranking the alternatives. Its simplicity and
taking into consideration both ideal and anti-ideal solutions are the reasons of its popularity
(Behzadian et al., 2012). The inputs of TOPSIS are the weights obtained in the first step and
the expert group average generated in the third step. After the weighted normalization, both
ideal and negative ideal distances are calculated. The final ratios of each supplier are
obtained, respectively, 0.711, 1.096, 2.488, 0.041 and 0.707. According to the ranking of final
ratios (from biggest to smallest), the ranking of the suppliers is obtained as Supplier 3 (2.488)
> Supplier 2 (1.096) > Supplier 1 (0.711) > Supplier 5 (0.707) > Supplier 4 (0.041).
5.2.2 Debiased ranking.
First step: manipulating criteria weights
It was shown that decision makers’ choices are affected from the “loss aversion effect.” The
criteria weights verify this finding as the highest weight is the cost’s weight. As a debiasing
technique, it is proposed to distribute the difference of the first (w1 5 wmax) and the second
(w2) highest weight to other criteria proportionally to the initial weights. Each criterion’s
debiased weights are calculated via equation (1). wdi represents the debiased weight of each
criterion and wi represents the biased weight of each criterion in this equation:

wdi ¼ wi þ ½ðwi =ð1  wmax ÞÞ3ðwmax  w2 Þ (1)

Table 6 shows the result of this process and criteria’s debiased weights.
Second step: determining experts’ bias degrees
This step is added to the previous one to determine the expert’s bias degrees. In this step, two
factors are used to determine the degree of the experts: “expertise level” and “response time.”
As Korhonen et al. (2018) generated their problem with random integers, aforementioned
factors are randomly created with discrete simulation in MS Excel via equation (2):

Expert group average S1 S2 S3 S4 S5

1 – Quality 3.4 4 2.8 4.2 2.4

2 – Cost 3 2.8 2.8 3.2 3.2
3 – Delivery speed 3.2 3.6 3 3.2 2.8
4 – Green image 3.2 2.6 4 2.8 3.8
5 – Pollution control 3 2.6 4 2.8 3.8 Table 5.
6 – Worker safety 2.6 3.2 3.8 3 3.6 Expert group average

Initial weights Debiased weights

C1 0.204 0.239
C2 0.323 0.204
C3 0.076 0.089
C4 0.179 0.210 Table 6.
C5 0.142 0.167 Initial and debiased
C6 0.077 0.091 weights
JEIM ¼ IFðC2 < 0:2; 1; IFðC2 < 0:4; 2; IFðC2 < 0:6; 3; IFðC2 < 0:8; 4; 5ÞÞÞÞ (2)

The decision rules about weighting according to the “expertise level” and “response time” are
as follows:
(1) The bigger expertise level, the bias will be lower. The expert’s weight will be higher.
(2) The lower response time, the bias will be higher. The expert’s weight will be lower.
Then, expertise level and response time degrees are normalized with dividing the degrees to
total degree. In the last step, as it can be seen in the last column of Table 7, total bias is
determined according to the average of both normalized levels.
Third step: obtaining supplier evaluations
It is aimed to take the SQ effect into consideration in this step. In accordance with the SQ
effect, there is an assumption that the decision makers have a tendency to stick to a given
default option (Geng, 2016). Due to the lack of information about the previous decisions of the
experts, the SQ effect is derived from expert opinions indirectly. An expert’s total score given
for each supplier is calculated and assumed that the biggest one is the previous supplier. To
add the SQ effect, it is experienced to reduce this supplier’s degrees one less. Because of the
highest total score of the second expert is 19 (fourth supplier), the degrees of the fourth
supplier’s column reduced one less and transformed to next column called “4 (debiased)”
(Table 8).
Fourth step: generating group decision
To generate group decision, experts’ decisions are weighted according to the bias degree
shown in the last column of Table 7. As a result, Table 9 is generated to be the input for
TOPSIS.
Fifth step: ranking suppliers with Technique for Order-Preference by Similarity to Ideal
Solution

Equally
Expert Expertise Response Response time Normalized Normalized weighted
No level time transformed expertise level response time Total bias

1 4 5 1 0.267 0.067 0.167

2 2 3 3 0.133 0.200 0.167
3 4 4 2 0.267 0.133 0.200
Table 7. 4 3 1 5 0.200 0.333 0.267
Bias generation 5 2 2 4 0.133 0.267 0.200
procedure TOTAL 15 15 1 1 1

2nd expert 1 2 3 4 4 (debiased) 5

1 – Quality 3 4 2 5 4 1
2 – Cost 2 2 2 4 3 3
3 – Delivery speed 3 4 2 3 2 3
4 – Green image 3 2 5 3 2 3
5 – Pollution control 3 3 4 2 1 3
Table 8. 6 – Worker safety 2 3 3 2 1 3
SQ effect reduction TOTAL 16 18 18 19 13 16
 Table 9 related to the expert group is used for TOPSIS input. The ratios obtained from the Cognitive
TOPSIS result of each supplier’s are shown in Table 10 compared with the biased version. biases in
According to the final ratios (from biggest to smallest), the ranking of the suppliers is
obtained as Supplier 3 > Supplier 1 > Supplier 2 > Supplier 5 > Supplier 4.
MCDM
Haeri and Rezaei (2019) obtained Supplier 3 > Supplier 5 > Supplier 2 > Supplier methods
1 > Supplier 4 ranking using an integrated gray relational degrees using an interval analysis
approach. As it can be seen in Table 10, for the biased and debiased versions, the best and the
worst suppliers did not change, but the ranking is changed. Supplier 5’s rank decreased from
second to fourth, and vice versa. MCDM techniques are used for not only choosing the best
alternative, but also ranking and classifying the alternatives. The bias effect gains more
important in such cases. Additionally, in another implementation, the best choice can be
changed according to the bias degree.

6. Conclusions and discussion/implications

To reduce cognitive biases that affect the decision-making process is the main subject of this
study. More specifically, the consequence of framing effect, loss aversion and SQ-type
cognitive biases on a sustainable supplier selection problem is studied. Moreover, it is aimed
to develop two pilot filters embedded in MCDM techniques as a debiasing strategy to
overcome the aforementioned biases. The first filter is related with framing effect, and the
second one is related with both loss aversion and SQ bias. In contrast to this literature, as a
novel approach, a manipulation of weights and a forecast of expected bias is proposed. The
result is also interesting because it is so simple and easily applicable.
This issue plays an important role in the BDT. As Morton and Fasolo (2009) state, a
psychologically aware MCDA practitioner plays a crucial role both in identifying the
necessary questions to ask and explain the gaps between the model and intuitive judgement.
As a result of this study, it is shown that biases effect our decisions. This finding has
important implications. This novel strategy generated in this study will be insightful for both
researchers and academicians working in MCDM and psychology. Biased decision-making
contributes to problems not only in private life but also in policy, medicine, business,

Expert group weighted sum S1 S2 S3 S4 S5

1 – Quality 3.437 3.904 2.302 3.937 2.035

2 – Cost 3.036 2.903 2.436 2.936 2.769
3 – Delivery speed 3.103 3.537 2.569 2.936 2.336
4 – Green image 3.270 2.569 3.504 2.703 3.537 Table 9.
5 – Pollution control 3.103 2.636 3.504 2.769 3.37 Expert group
6 – Worker safety 2.569 3.103 3.370 2.869 3.203 weighted sum

TOPSIS ratio Biased Debiased

S1 0.711 1.221
S2 1.096 1.178
S3 2.488 1.436
S4 0.041 0.011 Table 10.
S5 0.707 0.690 AHP-based TOPSIS
Ranking S3 > S2 > S1 > S5 > S4 S3 > S1 > S2 > S5 > S4 results
JEIM education and law (Morewedge et al., 2015). Debiasing studies are crucial to improve the
decision-making quality and performance (Cheng et al., 2014).
Unfortunately, there are limitations of this study. First of all, this study is an extended
version of a proceeding, and because of the time and resource limitations, data are obtained
randomly. To make experiments to verify the proposed methodology is one of the future
direction of this study. It is seen that the best practices are employed by decision and risk
analysts in debiasing, but few of them have been tested experimentally (Montibeller and
Winterfeldt, 2015). Borrero and Henao (2017) suggest that a totally rational decision-making
process – and therefore, a completely objective, analytical, tool-supported and evidence-based
management might be a utopic endeavor. In this experiments to reduce the SQ effect, to take
the last ten choices of each experts will be a more objective way. Also expertise level and
response time can be monitored quantitatively. As Morton and Fasolo (2009), sensitivity
analysis is omitted in this study, but the next step of this research is making sensitivity
analysis.
As a future direction, this novel debiasing strategy can be developed by adding other
types of biases that affect decisions. Indeed, not only the effects of some biases but also the
joint effect of the bias groups need to be studied. It is also interesting to study other MCDM
methods such as ELECTRE, PROMETHEE, ARAS, etc. By this way, the ultimate goal is to
find the most resistant technique to biases can be found. The results of these kind of studies
will show us which MCDM technique is more resistant to which cognitive bias. As a
consequence, before evaluation, we can test ourselves according to inherent biases and
choose the MCDM techniques according to this result and hybridize them. Also, this
technique can be applied to other decision problems. Especially in health-related decisions, it
is highly important. The other implementation area is ranking the nominees who applied a
job. In the literature, it was shown that there are some moderators such as decision makers’
self-esteem, neuroticism, mode of processing and gender. As an example, the framing effect
regarding risky products was found larger for those in a negative mood, and women exhibit
greater framing effects than men (Fagley et al., 2010). Location selection is another kind of
decision problem. Another avenue for future research relates to the gaps in the literature
about the antecedents of cognitive biases like individual differences (such as gender) and
emotions (such as such as apprehension, fear or anxiety) (Fagley et al., 2010). To reduce the
bias, understanding the underlying psychological process is crucial (Lee, 2019). As decision-
making is a complex activity that is influenced by many factors (Korhonen et al., 2018),
cognitive biases’ antecedents need to be analyzed. Considerable empirical and experimental
evidence in the behavioral psychology literature testifies to the existence of the bias in human
decision-making. As the BDT is a sub-discipline of psychology (Morton and Fasolo, 2009) and
according to the functional magnetic resonance imaging (fMRI) studies, activities of emotion-
related brain regions, such as the amygdala and the orbitofrontal cortex (Osmont et al., 2015),
biases in decision-making need to be handled in an interdisciplinary perspective.

References
Almashat, S., Ayotte, B., Edelstein, B. and Margrett, J. (2008), “Framing effect debiasing in medical
decision making”, Patient Education and Counseling, Vol. 71, pp. 102-107.
Arabsheybani, A., Paydar, M.M. and Safaei, A.S. (2018), “An integrated fuzzy MOORA method and
FMEA technique for sustainable supplier selection considering quantity discounts and
supplier’s risk”, Journal of Cleaner Production, Vol. 190, pp. 577-591.
Bazerman, M.H. and Tenbrunsel, A. (1998), “The role of social context on decisions: integrating social
cognition and behavioral decision research”, Basic and Applied Social Psychology, Vol. 20 No. 1,
pp. 87-91.
Behzadian, M., Otaghsara, S.K., Yazdani, M. and Ignatius, J. (2012), “A state-of the-art survey of Cognitive
TOPSIS applications”, Expert Systems with Applications, Vol. 39, pp. 13051-13069.
biases in
Borrero, S. and Henao, F. (2017), “Can managers be really objective? Bias in multicriteria decision
analysis”, Academy of Strategic Management Journal, Vol. 16 No. 1, pp. 244-259.
MCDM
Braverman, J.A. and Blumenthal-Barby, J.S. (2012), “Assessment of the sunk-cost effect in clinical
methods
decision-making”, Social Science and Medicine, Vol. 75, pp. 186-192.
Buchanan, J.T. and Corner, J.L. (1997), “The effects of anchoring in interactive MCDM solution
methods”, Computers Ops Res, Vol. 24 No. 10, pp. 907-918.
Cheng, F. and Wu, C. (2010), “Debiasing the framing effect: the effect of warning and involvement”,
Decision Support Systems, Vol. 49, pp. 328-334.
Cheng, F.F., Wu, C.S. and Lin, H.H. (2014), “Reducing the influence of framing on internet consumers’
decisions: the role of elaboration”, Computers in Human Behavior, Vol. 37, pp. 56-63.
Croskerry, P., Singha, G. and Mamede, S. (2013a), “Cognitive debiasing 1: origins of bias and theory of
debiasing”, BMJ Qual Saf, Vol. 22, pp. 58-64.
Croskerry, P., Singha, G. and Mamede, S. (2013b), “Cognitive debiasing 2: impediments to and
strategies for change”, BMJ Qual Saf, Vol. 22, pp. 65-72.
Deniz, N. (2018), “Supplier selection problem in the behavioral operations management context”, Proceedings
of XIII Balkan Conference on Operational Research, held at Belgrade, Serbia during May 25-28, p. 47,
available at: http://balcor2018.fon.bg.ac.rs/download/BALCOR2018BookOfAbstracts.pdf.
Fagley, N.S., Coleman, J.G. and Simon, A.F. (2010), “Effects of framing, perspective taking, and
perspective (affective focus) on choice”, Personality and Individual Differences, Vol. 48,
pp. 264-269.
Frydman, C. and Rangel, A. (2014), “Debiasing the disposition effect by reducing the saliency
ofinformation about a stock’s purchase price”, Journal of Economic Behavior and Organization,
Vol. 107, pp. 541-552.
Geng, S. (2016), “Decision time, consideration time, and status quo bias”, Economic Inquiry, Vol. 54
No. 1, pp. 433-449.
Guo, L., Trueblood, J.S. and Diederich, A. (2017), “Thinking Fast increases framing effects in risky
decision making”, Psychological Science, Vol. 28 No. 4, pp. 530-543.
Haeri, S. and Rezaei, J. (2019), “A grey-based green supplier selection model for uncertain
environments”, Journal of Cleaner Production, Vol. 221, pp. 768-784.
Haita-Falah, C. (2017), “Sunk-cost fallacy and cognitive ability in individual decision-making”, Journal
of Economic Psychology, Vol. 58, pp. 44-59.
Huang, H.H., Su, H.J. and Chang, C.J. (2015), “The moderating effect of a no-choice option on risky
framing effect”, Asia Pacific Management Review, Vol. 20, pp. 18-23.
Hwang, C.L. and Yoon, K. (1981), Multiple Attribute Decision Making: Methods and Applications,
Springer-Verlag, New York.
Kahneman, D. (2013), Thinking Fast and Slow, Farrar, Straus and Giroux.
Kahneman, D. and Tversky, A. (1979), “Prospect theory: an analysis of decision under risk”,
Econometrica, Vol. 47 No. 2, pp. 263-292.
Kim, J., Kim, J. and Marshall, R. (2014), “Search for the underlying mechanism of framing effects in
multi-alternative and multi-attribute decision situations”, Journal of Business Research, Vol. 67,
pp. 378-385.
Korhonen, P. and Wallenius, J. (1996), “Letter to the editor behavioural issues in MCDM: neglected
research questions”, Journal of Multi-Criteria Decision Analysis, Vol. 5, pp. 178-182.
Korhonen, P.J., Malo, P., Pajala, T., Ravaja, N., Somervuori, O. and Wallenius, J. (2018), “Decision
Support Context matters: the impact of product type, emotional attachment and information
overload on choice quality”, European Journal of Operational Research, Vol. 264, pp. 270-279.
JEIM Lee, K.K. (2019) “An indirect debiasing method: priming a target attribute reduces judgmental biases
in likelihood estimations”, PloS One, Vol. 14 No. 3, pp. 1-13.
Ma, W., Luo, X. and Jiang, Y. (2017), “Multicriteria decision making with cognitive limitations: a DS/
AHP-based approach”, International Journal of Intelligent Systems, Vol. 32, pp. 686-721.
Miu, A.C. and Crişan, L.G. (2011), “Cognitive reappraisal reduces the susceptibility to the framing
effect in economic decision making”, Personality and Individual Differences, Vol. 51,
pp. 478-482.
Montibeller, G. and Winterfeldt, D. (2015), “Cognitive and motivational biases in decision and risk
analysis”, Risk Analysis, Vol. 35 No. 7, pp. 1230-1251.
Morewedge, C.K., Yoon, H., Scopelliti, I., Symborski, C.W., Korris, J.H. and Kassam, K.S. (2015),
“Debiasing decisions: improved decision making with a single training intervention”,
Behavioral and Brain Sciences, Vol. 2 No. 1, pp. 129-140.
Morton, A. and Fasolo, B. (2009). “Behavioural decision theory for multi-criteria decision analysis: a
guided tour”, Journal of the Operational Research Society, Vol. 60 No. 2, pp. 268-275.
Osmont, A., Cassotti, M., Agogue, M., Houde, O. and Moutier, S. (2015), “Does ambiguity aversion
influence the framing effect during decision making?”, Psychonomic Bulletin and Review,
Vol. 22, pp. 572-577.
Pandey, R. and Jessica, V.M. (2018), “Measuring behavioral biases affecting real estate investment
decisions in India: using IRT”, International Journal of Housing Markets and Analysis, Vol. 11
No. 4, pp. 648-668.
Perfecto, H., Nelson, L.D. and Moore, D.A. (2018), “The category size bias: a mere misunderstanding”,
Judgment and Decision Making, Vol. 13 No. 2, pp. 170-184.
Qian, X., Fang, S.C., Huang, M. and Wang, X. (2019), “Winner determination of loss-averse buyers with
incomplete information in multiattribute reverse auctions for clean energy device procurement”,
Energy, Vol. 177, pp. 276-292.
Ryan, A., Duignan, S., Kenny, D. and McMahon, C.J. (2018), “Decision making in paediatric
cardiology. Are we prone to heuristics, biases and traps?”, Pediatric Cardiology, Vol. 39,
pp. 160-167.
Saaty, R.W. (2016), Decision Making in Complex Environments- the Analytic Network Process (ANP)
for Dependence and Feedback, Super Decisions, Pittsburgh.
Sellier, A.L., Scopelliti, I. and Morewedge, C.K. (2019), “Debiasing training improves decision making
in the field”, Psychological Science, Vol. 30 No. 9, pp. 1371-1379.
Shu, S. (2018), “Debias human judgment with a common decision bias: an experimental study of the
anchoring effect in hiring decision-making”, A Dissertation Thesis of PhD, University of
Minnesota Digital Conservancy, available at: http://hdl.handle.net/11299/199016.
Soman, D. and Liu, M.W. (2011), “Debiasing or rebiasing? Moderating the illusion of delayed
incentives”, Journal of Economic Psychology, Vol. 32, pp. 307-316.
 Pamucar, D., Puska, A. and Chatterjee, P. (2020), “Sustainable supplier selection in
Stevic, Z.,
healthcare industries using a new MCDM method: measurement of alternatives and ranking
according to COmpromise solution (MARCOS)”, Computers and Industrial Engineering,
Vol. 140, pp. 1-15.
Tversky, A. and Kahneman, D. (1974), “Judgment under uncertainty: heuristics and biases”, Science
New Series, Vol. 185 No. 4, pp. 1124-1131.
Walmsley, S. and Gilbey, A. (2016), “Cognitive biases in visual pilots’ weather-related decision
making”, Applied Cognitive Psychology, Vol. 30, pp. 532-543.
Zimmer, K., Fr€ohling, M. and Schultmann, F. (2016), “Sustainable supplier management – a review of
models supporting sustainable supplier selection, monitoring and development”, International
Journal of Production Research, Vol. 54 No. 5, pp. 1412-1442.
About the author Cognitive
Nurcan Deniz is an Assistant Professor Doctor of Operations Management and Marketing at Eskisehir
Osmangazi University, Turkey. She received her BS degrees from the Industrial Engineering and biases in
Chemical Engineering (Double Major) Department of the Eskisehir Osmangazi University, in 2005 and MCDM
2006. Also, she graduated from Business Administration Open Faculty in Anadolu University (2005). methods
Her MSc degrees are from Statistics (2009) and Industrial Engineering (2015). Her first PhD degree is
from Business Administration (2016). Her main research interests are in the areas of supply chain
management, service quality and lean production. She also continues her second PhD in Industrial
Engineering Department. Nurcan Deniz is the corresponding author and can be contacted at: ndeniz@
ogu.edu.tr

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com