You are on page 1of 29

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/318863473

Discovery Learning in Management Education:


Design and Case Analysis

Article in Organizational Behavior Teaching Review · August 2017


DOI: 10.1177/1052562917720710

CITATIONS READS

43 979

2 authors, including:

Daniel Druckman
George Mason University, Macquarie University (Sydney)
249 PUBLICATIONS 6,016 CITATIONS

SEE PROFILE

All content following this page was uploaded by Daniel Druckman on 26 February 2018.

The user has requested enhancement of the downloaded file.


XXX10.1177/1052562917720710Journal of Management EducationDruckman and Ebner
research-article2017

Article
Journal of Management Education
1­–28
Discovery Learning in © The Author(s) 2017
Reprints and permissions:
Management Education: sagepub.com/journalsPermissions.nav
DOI: 10.1177/1052562917720710
https://doi.org/10.1177/1052562917720710
Design and Case Analysis journals.sagepub.com/home/jme

Daniel Druckman1,2,3 and Noam Ebner4

Abstract
Two approaches to the process of guided discovery learning are compared
for their impacts on concept understanding. One, referred to as design,
emphasizes invention and draws on the simulation literature. The other,
referred to as case analysis, focuses on discovery and draws on the case-
based reasoning literature. Following a lecture on four cognitive-bias
concepts, management students were assigned randomly to a design, case
analysis, or lecture-only condition: A first experiment compared a design
with a lecture-only condition and a second experiment compared case
analysis with a lecture-only condition. Both design and case analysis students
understood the concepts better than lecture-only students. Designers in
the first experiment retained their understanding of the concepts better
than the case analysts in the second experiment. The impacts on learning
for design were similar to those obtained in earlier research where design
was compared with role-playing and a classroom lecture. Implications of the
findings for theory development and practice are discussed.

Keywords
case analysis, cognitive biases, concept learning, design, discovery learning,
management education

1George Mason University, Potomac, MD, USA


2Macquarie University, Sydney, New South Wales, Australia
3University of Queensland, Brisbane, Queensland, Australia
4Creighton University, Omaha, NE, USA

Corresponding Author:
Daniel Druckman, George Mason University, 10509 Gainsborough Road, Potomac, MD
20854, USA.
Email: dandruckman@yahoo.com
2 Journal of Management Education 00(0)

Introduction
In management education, the lecture-focused classroom has long since
given way to pedagogical approaches including more active student learning.
The case study method is among the most commonly encountered (see, e.g.,
Gamble & Jelley, 2014). Moving beyond engaging students in the type of
critical thinking and intellectual involvement this method incurs, other meth-
ods engage students in more active and personal encounters with manage-
ment principles and situations. The value of experiential learning for business
education is well recognized (see, e.g., Bevan & Kipka, 2012; McCarthy &
McCarthy, 2006), with much of the attention focused on simulation games,
either as in-class activities, or as part of wider competitions (see, e.g., Faria,
Hutchinson, Wellington, & Gold, 2009; Salas, Wildman, & Piccolo, 2009).
In this article, we suggest a new avenue or line of methods that engages
adult students actively in designing educational material with the goal of
achieving a deeper understanding of the material itself. Following the trend
toward discovery learning (DL), design encourages students to actively par-
ticipate in the learning process. Heeding the call stressing the need for research
comparing management teaching methods, rather than research focused on
individual methods (Burgoyne & Cooper, 1976), our research compares sev-
eral different teaching methods used in management education. These meth-
ods include the familiar classroom approaches of lecture and case studies as
well as the new method of design. Focusing attention on opportunities to
broaden opportunities for adult learning (Merriam, Caffarella, & Baumgartner,
2006), we do research on our own teaching practices (Elsner, 1991).
The meta-analysis conducted by Alfieri, Brooks, Aldrich, and Tenenbaum
(2011) suggests that the goal of achieving optimal concept learning can be
realized if at least one of the following support tools are provided:

a) guided tasks that have scaffolding in place to assist learners, b) tasks that
require learners to explain their own ideas, ensuring that these ideas are
accurate by providing timely feedback, and c) tasks that provide worked
examples of how to succeed in the task. (p. 13)

The key to DL may be found in the assistance that is provided by instructors.


Referred to in the literature as modeling or coaching, the assistance may
include execution of parts of the task in the form of worked problems, hand-
outs, and feedback, or props, hints, and cues (see Chi, 2009). These are ele-
ments of a larger process of scaffolding where teacher-assisted learning gives
way to student responsibility for learning as described by Wood, Bruner, and
Ross (1976) and Roehler and Cantlon (1997), as well as Berk and Winsler
Druckman and Ebner 3

(1995) on children’s learning. They are captured in the two types of guided
tasks compared in the experiments discussed below.
The key contributions made by this article are the following: (a) to distin-
guish between two modes of DL, design and case analysis, and review the
relevant literature on each approach; (b) to compare the impacts of each mode
with a classroom lecture on concept understanding; and (c) to compare the
impacts of the two DL modes. By comparing learning approaches, we extend
the literature on design/role-play (Druckman & Ebner, 2008) and case study/
lecture comparisons (Thistlethwaite et al., 2012). These comparisons move
the teaching literature forward from treating each approach in terms of its
own learning silo. We also situate the comparison in a new context, manage-
rial decision making, and use different concepts, cognitive biases. These fea-
tures contribute to issues of the extent to which designer effects generalize
across contexts. We also assess the value added to concept understanding by
following a lecture with a design or case analysis task. Implications for a
combination of the methods will be addressed in the Discussion section.
This article describes two experiments that compare these methods. We
begin with a discussion of the theoretical background for the approaches and
review earlier research to provide a context for the hypotheses addressed by
the experiments.

Literature Review
Discovery Learning
DL refers to tasks where the target information must be discovered by the
learner. The learner is at the center of the learning process. This idea has roots
in Whitehead’s (1929) process learning philosophy of education. For him,
learning is “a process of discovery, a process of becoming used to curious
thoughts, of shaping questions of seeking for answers, of devising new expe-
riences . . . ” (p. 32; see also Allan’s [2012] elaboration of Whitehead’s [1929]
process philosophy of learning). It also surfaces in Knowles, Holton, and
Swanson’s (2011) treatment of the adult learner, particularly with regard to a
need to be self-directing. This need is fulfilled when teachers expose students
to new learning possibilities and provide the conditions that are conducive to
active involvement with the material to be learned.
This approach to learning gained momentum with Bruner’s (1961) article.
Based on a constructivist approach to learning, he stressed the advantages of
active participation in the learning process through self-guided comprehen-
sion. But, he also cautioned that discovery is less likely to occur without
some knowledge of the domain in question. Building on this caution, Mayer
4 Journal of Management Education 00(0)

(2004) argued against unassisted discovery. His literature review led him to
conclude that unassisted discovery without structure is unlikely to improve
learning outcomes. Others have shown that more explicit instruction or
directed guidance is essential for effective DL (e.g., Clark, 2009; Rosenshine,
2009). The importance of direction is emphasized as well in Whitehead’s
(1929) approach to process learning (see also Allan, 2012, for further devel-
opment of this idea). Thus, learners need guidance or support tools to be
effective discoverers.
These studies are examples of an extensive literature on learning tech-
niques. That literature was evaluated in two meta-analyses conducted by
Alfieri et al (2011). The first analysis compared the effects of unassisted DL
conditions with explicit instruction in performing the learning tasks (a total of
580 comparisons from 108 studies). DL exercises supported by explicit
instruction was more beneficial than unassisted discovery tasks: The differ-
ence between these conditions turns on the extent to which assisted instruc-
tion (vs. self-instruction) is provided. The second analysis compared the
effects of enhanced DL approaches with a variety of instructional forms
including unassisted discovery and explicit instruction (a total of 360 com-
parisons from 56 studies). The enhanced discovery condition consisted of
guidance (generation, elicited explanation, or guided discovery) in perform-
ing the task. This condition resulted in better learning than any of the com-
parisons, especially for adult learners. The difference between “explicit” and
“enhanced” discovery is more a matter of degree or extent of guidance than
of the type of learning. The comparison conditions for the two analyses also
differed, with unassisted DL being used in the first, and a variety of instruc-
tional methods being used in the second, analyses.
The studies examined in the meta-analyses focused primarily on concept
learning in formal education contexts. This kind of learning is distinguished
from skills training or from what may be regarded as “everyday learning.”
Concept learning requires engagement in the learning task but also the chal-
lenge of constructing ideas that go beyond the presented information (to
elaborate, predict, and reflect; Chi, 2009). These tasks encourage learners to
strive for higher order objectives by being analytical, that is, engaging with
the material as analysts (Fletcher, 2009). The experiments reported in this
article are designed to evaluate the impacts of three modes of learning on
concept understanding.
Issues concerning the amount and type of guidance needed for effective
learning were addressed in a debate between Kirschner, Sweller, and Clark
(2006) and Schmidt, Loyens, van Gog, and Paas (2007). The former group
reviewed the evidence on the amount of guidance needed for effective learn-
ing, showing that minimal guidance, including various DL approaches, were
Druckman and Ebner 5

generally ineffective. The latter group concurred with this conclusion but
took exception with regard to problem-based learning (PBL) approaches,
considered by Kirschner and his colleagues to involve minimal guidance.
According to Schmidt and his colleagues, PBL allows for the flexible adapta-
tion of guidance, placing it perhaps in the middle of a continuum ranging
from maximal to minimal guidance. Furthermore, PBL approaches are tai-
lored to collaborative or group learning which was largely overlooked in the
Kirschner et al. (2006) review. PBL approaches are also helpful for applying
theoretical constructs to real-world problems. In these ways, PBL is similar
to our experimental conditions. In these, we ask learners to apply theoretical
constructs in collaborative exercises that take two forms, one involving
invention, the other consisting of discovery.
The guided tasks we explore consist of design and case analysis. Design
utilizes guided invention, by which we mean creating an original format and
narrative which may take the form of a script for an artistic performance or
instructions for role-players participating in an interactive simulation. Case
analysis consists of guided discovery, by which we mean using concepts to
provide explanations for managerial decision dilemmas described in a case.
The conceptual understanding gained from each of these tasks is compared
with the understanding that occurs following a classroom lecture on the con-
cepts. Understanding is assessed by survey questions administered immedi-
ately following the task (short-term understanding) and 1 week later
(long-term understanding). The longer term assessments capture the scaffold-
ing idea of fading (Puntambekar & Hubscher, 2005).
The lecture may be regarded as another form of assisted learning where
the instructor provides definitions, illustrates applications, and encourages
class discussion about the concepts. It precedes the design and case analysis
tasks but is also used as a standalone condition in the experiments. All three
conditions involve coaching (see Chi, 2009). A key distinction, however, is
the contrast between coaching with (design, case analysis) and without (lec-
ture) a performance to follow. The coaching occurs prior to the performance
in the design and case conditions.
The three learning modes can also be understood in terms of Whitehead’s
(1929) cyclical process learning model. Whitehead suggests that all learn-
ing must cycle through three stages, in an iterative manner. The first stage,
which Whitehead dubbed the stage of romance, involves engaging the
learner’s interest and curiosity. Students are introduced to concepts in a
general sense, and learn that these are important and meaningful, with far-
reaching implications—but do not yet consider or explore these implica-
tions. Their minds, to use Whitehead’s term, are fermenting, waiting for
new connections to set information in order. These are provided in
6 Journal of Management Education 00(0)

the second stage, referred to as precision. In this stage, learners gain an


analytical structure, a new system for organizing the information. This sys-
tem allows learners to fully comprehend the information that was intro-
duced to them in the stage of romance. The final stage is generalization, in
which learners, now holding conceptual information as well as a system for
applying it, can cast about to those areas that beckoned yet remained elu-
sive in the stage of romance, and understand them through applying their
newly gained concepts taking a systematic approach. Teaching any indi-
vidual learning process should cycle through these stages; learning, in gen-
eral, comprises multiple iterations. “Education,” Whitehouse says, “should
consist in a continual repetition of such cycles” (p. 19).
Lectures might be particularly helpful as a means of triggering the stage of
romance, particularly if they are designed to achieve this stage’s goals, rather
than aiming to provide students with a complete learning experience. Lectures
can also be helpful for providing an analytical structure in the stage of preci-
sion. Cases can be used as a preliminary method for identifying concepts
needing exploration in the stage of romance, and as an inductive method
through which students form an analytical structure in the stage of precision.
Design activities might be particularly oriented toward the third stage, that of
generalization, in which they encourage learners to synthesize new uses and
contexts for the information and analytical systems they have gained in the
previous stages. Moreover, as learners engage in the creative process, they
are likely to uncover new questions needing exploration—initiating a new
stage of romance.
Allan discusses the transition from the abstractions of Whitehead’s (2012)
second stage of precision to active application of the concepts provided by
the “freedom generalization (Whitehead’s third stage) offers a person” (p.
24). Lectures can be construed as the abstract learning part of the cycle while
design and cases are the more active elements. The design experience applies
the concepts in an inventive way; the case experience applies them in a more
inductive discovery mode. Implications of our findings for cyclical learning
is addressed in the Discussion section. We turn now to a discussion of each of
the three modes. The key terms used to discuss them in this article are
described in Table 1.

Classroom-Based Lectures
Lectures have been a part of the teacher’s activity from times immemorial. As
Bligh (1998) describes, our generation is no exception, and lectures seem to
defy the changing landscape of educational research and technology to
remain the most common method for teaching adults.
Druckman and Ebner 7

Table 1. Descriptions of Key Terms.

Term Description
Case analysis Learners are provided a case study and engage in a process
of applying specific concepts and frameworks—such as the
cognitive bias concepts used in these experiments—to aid
in discovering key factors at play.
Discovery A pedagogical approach centered around exposing learners
learning to new learning possibilities and offering them opportunities
to uncover the target information on their own.
Design A method of instruction in which the learner is asked to
actively create an artifact (e.g., an exercise, role material for
a simulation, a case study) of their own, rather than engage
with one created and/or provided by the teacher.
Fading/retention The extent to which understanding of concepts persists after
the learner has gained them through a learning activity.
Guided invention Learners are set to engage in a creative task, such as design,
with the teacher providing them a set of guidelines to follow
as they navigate the task.
Guided discovery Learners are asked to engage in an analytical task, such as
case analysis, with the teacher providing them a set of
guidelines to follow as they navigate the task.
Lecture A range of classroom activities in which all or most of the
information originates with the teacher and flows to the
students, largely from a teacher-designed plan.

Lectures—sessions in which speakers expound, more or less continu-


ously—on a topic they wish the audience to learn about, are utilized to
achieve four different types of pedagogical goals: information acquisition,
thought promotion, attitudinal change, and behavioral skills. However, the
efficacy of lectures for achieving these goals has not been proven (Bligh,
1998). They may be less effective than other methods as shown by the recent
meta-analysis conducted by Freeman et al. (2014). They reviewed 225 stud-
ies dealing with the efficacy of stand-and-deliver lectures on science, tech-
nology, engineering, and math topics. A key finding was that students taking
lecture-based sources were considerably more likely, calculated as 150%, to
fail them than students taking courses with at least some degree of active
learning.
Given these results, it is not surprising to encounter methods touted as
improvements on traditional lectures. Still it is important to recognize that
academic lectures differ one from another, on a number of dimensions, and do
not necessarily match Bligh’s unidirectional (teacher to student) depiction. As
8 Journal of Management Education 00(0)

a category, they involve a “sage on a stage” imparting wisdom in a fashion that


is both unidirectional in terms of communication as well as “take-it-or-leave
it” in terms of style; they take the form of sheer impartation of information,
without necessarily aiming to inspire, encourage, trigger, or engage.
A wider view of the lecture method would recognize other approaches to
lecturing, even when limiting lectures to a modality in which an entire class
engages with their teacher. Although the primary function of lectures is to
provide information, they can also pose challenging questions for students to
consider as the lecturer is talking. The teacher can tell stories, embedding
information as a moral to be learned or as a behavior to beware. They can
focus narrowly on a topic, or discuss that topic as it is embedded in a histori-
cal perspective or an interdisciplinary concern. As a form of active learning,
lectures can involve students in activities that transcend the model of absor-
bent minds, whether through Socratic or other modes of questioning and dis-
cussion. Lectures can rise above discussion, to involve higher order cognitive
activity such as comparing, analysis, and integration (Bain, 2004; see also
Knowles, 1984). Bain (2004) suggests that the best lecturers include ele-
ments of critical learning in their approach. Rather than directing information
at participants, they guide students toward the information and motivate them
to uncover it.
Bain’s (2004) survey of excellent teachers and their practices concluded
that lectures cannot be the only tool in a teacher’s arsenal. Lectures, he sug-
gests, serve best to raise questions and direct student attention at particular
issues, but these questions and issues need then to be addressed by other ele-
ments in the learning environment. Nonetheless, lectures remain the pre-
ferred approach to teaching despite evidence that questions their efficacy
(Bligh, 1998). There is, however, a discernable trend in the direction of
adopting a wider set of teaching methods and activities. We discuss two of
these alternatives in the sections to follow.

Design
Design is a form of DL that emphasizes invention. Design learners are
encouraged to represent concepts in the form of a scenario, script, or narra-
tive. The product students design might be the information required for a
role-play, the script and direction for a theatrical performance, or the text for
a speech to be given by others. An example of the design process is provided
by research conducted on uses of simulation for teaching negotiation
(Druckman & Ebner, 2008).
The learning advantages of design have been demonstrated in a series of
experiments in which student designers were compared with students who
Druckman and Ebner 9

conducted role-plays and with those who only listened to a lecture. Design
and role-play are both forms of assisted DL, with an emphasis on either
invention (design) or experience (role-play).1
The findings showed that the design method enjoyed significant advan-
tages over the role-playing method. Design was found to be better in terms of
concept learning as well as for the retention of this learning over time, which
is referred to as fading in the scaffolding literature. Furthermore, design pro-
duced stronger improvements in student interest and motivation than role-
play participation. This finding is particularly interesting, given that earlier
research had shown that the primary advantage of role-playing simulations
(compared with other learning approaches) is enhanced motivation (e.g.,
Ellington, Fowlie, & Gordon, 1998).
Another finding, suggesting differences between types of DL, bears on a
key feature of design—learning relationships among concepts. In the experi-
ment discussed above, design enhanced this type of learning more than con-
ducting role plays or listening only to a lecture. Negotiation concepts exist
within a web of relationships and interactions with other concepts. Finding
the best way to teach a single concept is surely important; being able to effec-
tively teach students the relationships among different concepts is a more
elusive goal, and it appeared that design tasks provided a key to this kind of
learning (Druckman & Ebner, 2008).
Features of design thought to contribute to learning include concreteness,
creativity, and involvement (Crookall, 1995). These features highlight dis-
covery through invention. But, perhaps the key pedagogical advantage of
design is learning about relations among different concepts (Greenblat, 1981)
or developing an understanding of a system of relationships among them,
referred to by Duke (1974) as concept maps. These insights may be a basis
for a theory of design that emphasizes synthetic learning.
These results support earlier findings from a more general literature on
assisted DL. As we note above, the Alfieri et al. (2011) meta-analysis
shows advantages for assisted DL. The comparison between design and
lecture, referred to as explicit instruction, reinforces these findings. In
addition, it contributes another set of findings: Design as a guided inven-
tion process is shown to have both learning and motivational advantages
over role-play as a guided experiential form of DL. One implication is that
some methods of assisted DL (e.g., design) are more effective than other
methods (e.g., role-play). In this article, we extend the portfolio of learn-
ing modes by comparing design-based learning with case analysis (and
lectures). We also change the design task from scenario development for
role-plays (the task assigned in previous experiments) to script writing for
theatrical performances.
10 Journal of Management Education 00(0)

In a first experiment, we evaluate hypotheses that compare short-term and


retention learning through design with learning the same concepts through
explicit instruction. Retention assessments address the scaffolding idea of
fading: To what extent are the understandings retained after the initial task is
performed? The hypothesized comparisons are made with different concepts
(cognitive biases), in a different setting (a course on managerial decision
making in a business school), and using a different design exercise (authoring
scripts). The results will shed light on the robustness of design impacts on
learning. They will also provide a basis for theory development, to be
addressed further in the Discussion section.

Case Analysis
Proponents of the case method suggest that it has many advantages over tradi-
tional classroom lecture–based learning (Kunselman & Johnson, 2004). As a
form of DL, the method has active components: Students need to bring their
knowledge to bear on a case, identify the core problems it presents, and iden-
tify key questions that need to be answered. Case analysis challenges them to
discover the way concepts are manifest in the case material with which they
are presented. This guided discovery process is different from the guided
invention process that occurs with design activities. Rather than challenging
students to devise cases in the form of scenarios or scripts, case analysis chal-
lenges them to discover the way concepts are manifest in the case material
with which they are presented.
The analysis process may be thought about as being similar to the work of
detectives after they have compiled the evidence relevant to a crime. They
comb the evidence for clues that suggest directions for solving the crime. This
is an inductive approach to problem solving akin to the way analysts search
for clues to understand the dynamics of an historical case. Both engage in a
search and discover activity. But the approach also has elements of deduction
when the detectives and analysts are guided by concepts. Further insights
about these processes come from the case-based reasoning (CBR) literature.
A key cognitive process involved in CBR is thinking by analogy (Kolodner,
1997). Case analysts search for correspondences between concepts and the
situation described in the case. The process includes the steps of retrieval of
concepts, searching for correspondences, inducing relationships between the
concepts and actions, and learning to abstract commonalties shared by the
concepts and the situation (Holyoak & Thagard, 1997). These steps facilitate
analogic transfer from a base or source to a target case. The base in our exper-
iments described below consist of examples of cognitive bias concepts used
in a lecture. The target is the new case used by students to perform analyses.
Druckman and Ebner 11

Correspondences between base (teacher case examples) and target (analysis


case) indicate the soundness of the transfer.
The learning that occurs is similar to the research methodology known as
enhanced case studies (Druckman, 2005). Both the classroom activity of
using concepts as tools for understanding cases and the research activity of
analyzing cases with theoretical concepts are inductive learning strategies
that feature reasoning by analogy. Both involve diagnosing situations through
the lenses of more general ideas (Gentner & Holyoak, 1997) and may be the
basis for a more general theory of inductive learning.
The shared features of CBR take a variety of forms when used in class-
room situations. In her examination of the epistemological bases for the dif-
ferent uses of cases, Kagan (1993) surveyed applications of the case method
from 1927 to 1991. The applications included learning pedagogical princi-
ples, critical incidents, developing problem-solving strategies, class evalua-
tion of student solutions to a problem, discussion-based active learning,
writing cases, scenario-based role-plays, and guided learning provided by
teachers. These differences often reflect preferred approaches to teaching a
practice (medicine, law, business, social science) and suggest “that teaching
is unique as a clinical profession in its attention to the elements of practice
that may be practitioner-specific (expressive)” (Kagan, 1993, p. 715). The
approach to case analysis taken in the experiments to follow also reflect a
particular practice goal, namely, to enhance understanding of a complex
organizational decision situation.
The contributions made by CBR researchers have increased our under-
standing of cognitive and motivational processes as well as the larger envi-
ronments in which cases have been used. Less attention has been paid to the
comparative evaluation of case analysis. Studies comparing the learning ben-
efits of the case method with those of lecture-based learning—in business
education and in other disciplines—have yielded mixed results. While some
studies indicate learning advantages associated with the case method, many
others tend to show no difference in student learning between the two teach-
ing methods. The data are, at best, inconclusive, and tend to indicate that the
case method shows motivational rather than learning gains (see Carrero,
Gomar, Penzo, & Rull, 2007; S. Moore, 1999; also see the meta-review con-
ducted by Thistlethwaite et al., 2012, in the area of health education).2 We
turn now to the set of hypotheses to be evaluated by the experiments.

Hypotheses
The experiments are intended to evaluate five hypotheses about the three learn-
ing approaches. Hypotheses 1 and 2 are derived from findings obtained in the
12 Journal of Management Education 00(0)

previous research on the use of design for learning negotiation concepts. They
address the value added by performing a design task following the lecture.

Hypothesis 1: A design process following a lecture increases short-term


understanding of the bias concepts more than a lecture without a design
process.
Hypothesis 2: A design process following a lecture increases longer term
understanding of the concepts more than a lecture without a design process.

Hypotheses 3 and 4 address the value added from performing a case analysis
following the lecture.

Hypothesis 3: A case analysis process following a lecture increases short-


term understanding of the bias concepts more than a lecture without case
analysis.
Hypothesis 4: A case analysis process following a lecture increase longer
term understanding of the concepts more than a lecture without the case
analysis.

A fifth hypothesis compares the learning impact of design, engaged in by


students in the first experiment, with case analysis using the same bias con-
cepts in a second experiment. Time spent on task was the same for both of
these exercises. Hypotheses 1 and 2, based on earlier findings, suggest that
design increases short- and long-term retention of concepts more than a class
lecture. Hypotheses 3 and 4 suggest that case analyses have advantages for
short-term and retention learning when compared with a lecture-only condi-
tion. Earlier results showing stronger effects for design–lecturer comparisons
than for case analysis–lecture comparisons suggest the following hypothesis.

Hypothesis 5: A design process following a lecture increases understand-


ing of the concepts more than a case study analysis following a lecture.

The experiments provide evidence for each of these hypotheses. Results


are presented following a description of the methods. The article concludes
with a discussion of implications of the findings for pedagogical theory and
practice as well as for further research.

Method
This section is divided into several parts. The setting and concepts to be
learned are discussed before turning to the design and procedures used in
each of the experiments.
Druckman and Ebner 13

Setting
The experiments were conducted with students in the 2014 and 2015 MBA
classes taught at the Melbourne Business School. The class topic was
Managerial Judgment. Students participated in a lecture that first discussed
judgment research in general and then addressed each of four cognitive biases:
overconfidence, framing, anchoring, and availability. In terms of the lecture
styles discussed earlier in this article, this lecture did not take the form of the
unidirectional, dispassionate impartation of information described by Bligh
(1998), but rather adopted the style of bidirectional interaction advocated by
Bain (2004), in which the teacher seeks not only to impart information but to
challenge—and motivate—students to uncover it for themselves. Initially,
each bias was defined and illustrated with examples supplemented by pictures
and occasional cartoons displayed with power point graphics. The examples
were often derived from experimental findings and formed the basis for class
discussion. The students were then asked to identify and consider manifesta-
tions of each bias in their own decision-making processes. Finally, they were
presented with possible remedies for each bias. For example, the availability
bias was defined in terms of its components, recency and vividness, and sup-
plemented with such remedies as “don’t trust your memory,” “keep records,”
and “beware of anecdotes.” The duration of the lecture was 1 hour.
Two experiments focused on learning about four cognitive biases, which
were the topics of a class lecture that preceded the exercises. These concepts
include overconfidence, framing, anchoring, and availability. They are
described in Table 2.

Experiments
Two experiments were conducted. The first is a comparison of a design task
following a lecture on the four bias concepts with a lecture-only condition.
Two bias concepts were used in the design task; the other two were presented
only during the class lecture. This experiment provides an evaluation of
learning with and without a design experience. The second experiment con-
sisted of a comparison of a case study analysis following a lecture on the four
bias concepts with a lecture-only condition. As in the first experiment, two
concepts were used in the analysis, while the other two concepts were pre-
sented only in the class lecture. In this section, we describe the design and
procedures used in each of the two experiments.

Experiment I
Experimental design. All students participated in a design task on two of the
four bias concepts presented in a lecture by the course instructor. The focal
14 Journal of Management Education 00(0)

Table 2. Four Cognitive Biases.

Bias Description Key source


Overconfidence People have more confidence in D. A. Moore and Healy
their judgments than is warranted. (2008)
Framing The way an issue is formulated sets Tversky and Kahneman
the frame that draws attention to (1981)
some of its features.
Anchoring People form a starting point based Tversky and Kahneman
on what they know—an anchor— (1974)
and adjust it to a more acceptable
value.
Availability People’s judgments are overly Taylor (1982)
influenced by readily available and
retrievable information

concepts were assigned randomly to partner dyads. Forty-eight students were


divided into two bias concepts conditions: framing/availability and overcon-
fidence/anchoring. A within-subjects experimental design was implemented
with students serving as their own controls. This design allowed for a direct
comparison of learning the focal (design task) and nonfocal (lecture-only)
concepts. Another advantage of this design is that it does not alter the class
plan, where all the concepts are taught to all the students. A between-sub-
jects design would entail exposing students to two of the four concepts only,
depending on their randomly assigned condition.
The analyses of the survey responses consisted primarily of repeated-mea-
sures comparisons using matched-pairs t tests for each of the dependent vari-
ables discussed in the next section.

Procedures. Following the lecture, the class was divided into two randomly
assigned groups, a framing/availability group and an overconfidence/anchor-
ing group. Work team pairings within each of these groups were also randomly
determined. All pairs were then given instructions for the designer task. The
task consisted of creating skits that could be used as a learning experience. The
skit was likened to a theatrical performance where students would be assigned
actor roles and interactions would unfold according to the script. Foremost in
their skit preparation, students were to incorporate the relevant cognitive biases
in a role-play context or story line. Furthermore, they were told that:

You will have one hour to design your skits, collaboratively in two-person
teams. With your partner, create two theatrical role-plays (skits), one that
demonstrates the over-confidence (or framing) bias, and one that demonstrates
Druckman and Ebner 15

the anchoring (or availability) bias as they might play out in a work setting.
Each script may run about five minutes.

The design process includes elements of assisted learning. Instruction is


provided about the concepts to be employed and instructions are given about
how to perform the task. Aspects of scaffolding captured by design include
hints, illustrative applications, diagnostic examples, peer support, and fading
(Puntambekar & Hubscher, 2005). It differs from scaffolding, however, with
regard to monitoring, calibrating, and providing feedback on performance
(Pea, 2004). Thus, design emphasizes the learning environment more than
assessment of learner progress.
At the completion of the design exercise, participants were given a survey
consisting of both closed- and open-ended questions dealing with under-
standing each of the four bias concepts (the two focal and two nonfocal con-
cepts). Following institutional review board guidelines, participation in the
survey was voluntary and anonymous. Each student was assigned an identi-
fication number which they would need to save. Responses to these questions
were the dependent variables. The survey is shown in the appendix. It was
administered again at the beginning of the next class, 1 week later, for assess-
ing learning retention. Students were asked to put their ID number on the
survey. At the completion of the retention survey, the students were debriefed.
The debriefing included information about the purpose of the exercise and
included findings obtained from earlier designer experiments.

Coding. The concept understanding question for each bias was presented on
a six-step scale with lower codes indicating more understanding. The question
about gains from the exercise was presented on a three-step scales including
“would have learned more (from a lecture-only),’ “about the same,” and “would
have learned less (from a lecture-only).” Two open-ended short essay questions
asked about the “meaning” of each bias and “identifying the bias in your work
as a manager.” The meaning question was coded as “correct,” “partially cor-
rect,” or “incorrect.” The bias identification question was coded as “insightful,”
partially insightful,” or “not insightful.” All essays were judged by two inde-
pendent coders. A calibration procedure was used for resolving differences.
The coders discussed reasons for the judgments and strove toward a consensus.
In rare instances, when consensus was not achieved, a third coder was brought
into the discussion and served to arbitrate the difference. Repeated-measures
analyses were performed on these data to assess each of the hypotheses.

Experiment II
Experimental design. All students participated in a case analysis task on
two of the four bias concepts presented in the lecture. The focal concepts were
16 Journal of Management Education 00(0)

assigned randomly to small work teams. Sixty-nine students were divided


into two bias concepts conditions: framing/availability and overconfidence/
anchoring. Fifteen teams worked for an hour on each pair of concepts (either
the framing/availability pair or the overconfidence/anchoring pair). A within-
subjects experimental design was implemented with students serving as their
own controls. This experimental design allowed for the same comparison
of learning advantages as in Experiment I. Furthermore, repeated measures
statistics were used to analyze the survey data. Different than the previous
experiment, however, the students performed analyses of a case study rather
than a design task.

Procedures. The experiment was conducted with students in the 2015


MBA class on Managerial Judgment. The lecture on the biases was the same
as that presented to the 2014 students participating in Experiment I. The same
random assignment procedure to the bias conditions and work teams was
also implemented. At this point, a different path was taken. Rather than mov-
ing toward a design task, these students were instructed to analyze a case
assigned for reading the previous week. A Harvard Business School case titled
“Man Jit Singh at Sony Entertainment Television” was used. This complex
multi-issue case was chosen primarily because it captured the purpose of the
course. This case highlighted a managerial judgment dilemma experienced
by the chief executive officer of Multi Screen Media Private Limited. The
dilemma concerned a performance evaluation issue of the relative weights to
be assigned performance (meeting financial and market share targets) versus
corporate values (accountability, empowerment, transparency, collaboration,
and innovation). His colleagues on the management committee were split on
the weights. Man Jit Singh was faced with the prospect of making a decision
that could affect the company’s competitive performance. Having read the
case prior to the class, students were now asked to apply the bias concepts
in an analysis. They were asked two questions: “Where in the case are these
biases likely to be operating?” and “What advice would you give?”
As with the design task, the case analysis process includes elements of
assisted learning. Instruction is provided about the concepts to be employed
and instructions are given about how to perform the task. Aspects of CBR
captured by the case analysis include inducing correspondences between the
bias concepts taught with examples in the lecture (base or source examples)
and the case situation (target case) as well as setting explicit practice goals.
Although analogistic thinking is not made explicit in task instructions, stu-
dents were free to use analogies as a tool for matching the concrete chal-
lenges of the target case with the source abstract concepts learned in the
lecture.
Druckman and Ebner 17

As in Experiment I, students were asked to complete the survey shown in a


Supplemental Appendix (all supplementary materials are available online at
http://jme.sagepub.com/content/by/supplemental-data). The survey was
administered again a week later during the next class followed by a debriefing
that included a discussion of both experiments. The survey data were used for
the analyses discussed in the next section. As in Experiment I, participation in
the survey was voluntary and anonymous. Coding was performed in the same
way as for Experiment I. Differences between coding judgments made on the
open-ended essays were calibrated by two independent coders.

Results
The results are organized by the five hypotheses.

Hypothesis 1: A design process following a lecture increases short-term


understanding of the concepts more than a lecture without a design
process.

A series of matched-pair t tests compared the impacts of the biases on self-


reported understanding of the concepts. Each of the design versus no design
comparisons was significant. For the comparisons involving the design con-
ditions of overconfidence and anchoring (lower means = more reported
understanding): Framing (M = 4.32) versus anchoring (M = 3.16; t = 3.09, p
< .003); overconfidence (M = 3.04 versus framing (M = 4.32; t = 3.73, p <
.001); overconfidence (3.04) versus availability (M = 4.64;t = 4.31, p <
.0001), and anchoring (M = 3.16) versus availability (M = 4.64; p < .0001). A
similar pattern occurred when framing and availability were the focal bases
in the design condition. All the t ratios were less than .001. None of the bias
comparisons within conditions (either design or control) approached signifi-
cance. Significant differences occur also for the short essay questions with
availability in the designer condition producing more accurate definitions of
the bias than both anchoring and overconfidence in the lecture-only condi-
tions (p < .004 and .007, respectively). These findings provide strong support
for Hypothesis 1 (see the means in Table 3).

Hypothesis 2: A design process following a lecture increases longer term


understanding of the concepts more than a lecture without a design
process.

A similar pattern of results occurred for the 1-week delay assessments with
seven of the eight design versus control comparisons producing significant t
18 Journal of Management Education 00(0)

Table 3. Means for Short-Term Understanding: Experiment I Pair Comparisons.

Comparisona Design Lecture only


Overconfidence/framing 3.04 4.32
Overconfidence/availability 3.04 4.64
Anchoring/framing 3.16 4.32
Anchoring/availability 3.16 4.64
Framing/overconfidence 2.33 3.81
Framing/anchoring 2.33 4.29
Availability/overconfidence 2.14 3.81
Availability/anchoring 2.14 4.29
aThe first bias in each pair comparison is the design condition; the second bias is the lecture-

only condition.

Table 4. Means for Retention Understanding: Experiment I Pair Comparisons.

Comparisona Design Lecture only


Overconfidence/framing 3.25 3.99
Overconfidence/availability 3.25 4.18
Anchoring/framing 3.21 3.99
Anchoring/availability 3.21 4.18
Framing/overconfidence 2.90 3.45
Framing/anchoring 2.90 3.32
Availability/overconfidence 2.68 3.45
Availability/anchoring 2.68 3.32
aThe first bias in each pair comparison is the design condition; the second bias is the lecture-

only condition.

ratios (with probabilities ranging from .05 to .008; see Table 4 for the means).
Significant differences were also obtained in the expected direction for the
short essays on definitions of the concepts. These findings provide strong sup-
port for Hypothesis 2.

Hypothesis 3: A case analysis process following a lecture increases short-


term understanding of the bias concepts more than a lecture without case
analysis (lecture-only).

Significant differences occurred for the comparisons between overconfi-


dence (case study) and availability (lecture-only; p < .03) and between anchor-
ing (case study) and availability (lecture-only; p < .008). A comparable pattern
Druckman and Ebner 19

Table 5. Means for Short-Term Understanding: Experiment II Pair Comparisons.

Comparisona Case analysis Lecture only


Overconfidence/framing 3 3.30
Overconfidence/availability 3 3.78
Anchoring/framing 2.83 3.30
Anchoring/availability 2.83 3.78
Framing/overconfidence 2.63 3.11
Framing/anchoring 2.63 3.30
Availability/overconfidence 2.82 3.11
Availability/anchoring 2.82 3.30
aThe first bias in each pair comparison is the case analysis condition; the second bias is the

lecture-only condition.

occurs for comparisons involving the framing bias: framing (case study) pro-
duced more understanding than the lecture-only concepts of overconfidence
(p < .08) and anchoring (p < .01). The difference between the availability (case
analysis) and overconfidence (lecture-only) biases were not significant; nor
was the difference between the availability (case analysis) and anchoring (lec-
ture-only) biases significant. However, the availability bias case analysis con-
dition did produce more accurate definitions than both the lecture-only
overconfidence (Ms = 1.30 and 1.70, respectively; p < .02) and anchoring
biases (Ms = 1.30 and 1.78, respectively; p < .02) conditions. These findings
provide partial supporting evidence for this hypothesis: case analysis enhances
learning on some of the bias concepts (see Table 5 for the means).

Hypothesis 4: A case analysis process following a lecture increase longer


term understanding of the concepts more than a lecture without the case
analysis.

The retention data show no significant differences among the means for
any of the case study versus lecture-only comparisons. The framing (M =
3.56) and availability (M = 3.44) case analyses produce similar understanding
to the overconfidence (M = 3.44) and anchoring (M = 3.22) lecture-only con-
ditions. A similar pattern of mean differences occurred when overconfidence
(M = 2.62) and anchoring (M = 2.77) were used for the case analyses (the
framing and availability means were both 2.85). The case analysis and lecture-
only conditions did not differ on the accuracy of the open-ended questions
about the definition of the bias concepts. Thus, these findings provide evi-
dence for this hypothesis: Case analyses do not enhance understanding of the
concepts more than when only a lecture is given (see Table 6 for the means).
20 Journal of Management Education 00(0)

Table 6. Means for Retention Understanding: Experiment II Pair Comparisons.

Comparisona Case analysis Lecture only


Overconfidence/framing 2.62 2.85
Overconfidence/availability 2.62 2.85
Anchoring/framing 2.77 2.85
Anchoring/availability 2.77 3.85
Framing/overconfidence 3.56 3.44
Framing/anchoring 3.56 3.22
Availability/overconfidence 3.44 3.44
Availability/anchoring 2.44 3.22
aThe first bias in each pair comparison is the case analysis condition; the second bias is the

lecture-only condition.

Hypothesis 5: A design process following a lecture increases understand-


ing of the concepts more than a case study analysis following a lecture.

A significant difference occurred between the design and case analysis


concept of availability. More understanding was shown when this concept
was used in a design than in a case analysis exercise (M [design] = 2.14; M
[case analysis] = 2.82; p < .07). The difference between these conditions for
the framing concept was in the predicted direction but not significant (M
[design] = 2.33; M [case analysis] = 2.63). The case analysis showed better
learning than the lecture-only condition used in the design experiment
(Experiment 1; p < .025 and p < .02 for overconfidence and anchoring,
respectively). No differences in understanding between the design and case
analysis conditions occurred when overconfidence and anchoring were used
as the focal biases for design or case analysis. The short essays on defining
the biases showed that the best definitions were provided for the availability
bias in both the design and case study conditions (means of 1.3 for both con-
ditions compared with means that ranged from 1.6 to 2.1 in the other condi-
tions). These findings provide evidence for this hypothesis: design enhances
learning more than case analysis on some of the bias concepts.

Discussion
In this section, we discuss the key contributions made by this study. One
contribution is the distinction between two modes of DL, design and case
analysis. Referred to as assisted DL (Alfieri et al., 2011), these modes were
used to enhance the understanding of social science concepts. The design
Druckman and Ebner 21

mode highlights invention by challenging students to create an original script


that incorporates the concepts. The case analysis mode encourages induction
by challenging students to search for clues in the case that illuminate the
concepts.
Both learning modes produced better understanding of the bias concepts
than a lecture. Designer effects were significant across the four concepts for
both immediate and retention learning. Case analysis effects were significant
for three of the four concepts for immediate learning, the exception being the
anchoring bias. Interestingly, direct design/case analysis comparisons for
short-term learning showed a significant difference in favor of design for
only the availability concept. However, the benefits of learning through case
analysis proved to affect short-term learning only, and did not hold up in the
retention assessments—in which design’s learning effects proved to be more
beneficial than case analysis. These results suggest that the understanding
produced by design, highlighting the invention process, is sustained better
over the course of a week than the understanding produced by the discovery
process involved in case analysis. A recommendation for teachers based on
these findings is to augment lectures with both design and case exercises.
However, given a choice, the design exercise may be the preferred route with
concepts for learning concepts similar to the cognitive biases explored in
these studies.
These activities are akin to the way scientists perform analysis: design
captures elements of designing experiments; case analysis is similar to impos-
ing theoretical concepts on cases referred to as an enhanced case study (see
Druckman, 2005, Chapter 6). This distinction, applied both to students and
scientists, suggests possible avenues for theory development.
One avenue moves us in the direction of a theory of design. The walk
begins with the idea of using abstract concepts in the creation of a narrative.
Traveling further down the avenue, the designer confronts the challenge of
embedding the concepts (cognitive biases) within the narrative or script. For
example, a framing bias is evident in a script about a political discussion on
the limits of refugee immigration to one’s country. Another corner on the
avenue is turned when the designer attempts to relate two or more bias con-
cepts. For example, the framing bias is complemented by an availability bias
when the immigration argument is developed from information found in
recent updates provided by favored political organizations. Approaching the
end of the avenue, the designer writes the narrative in a form that can be
implemented by performers or role-players. Illustrated by this journey are the
ideas of grounding abstractions, conceptual mapping, relations between con-
cepts, and communication. Together, these steps in the journey, taken by our
designers, are an exercise in synthesis. A theory of design may essentially be
22 Journal of Management Education 00(0)

a theory of synthetic thinking (see Dempsey, 2010, for ideas that lead to a
theory of synthetic learning).
Another avenue moves us in the direction of a theory of case analysis. This
trip begins with using abstract concepts for diagnosing cases. Traveling down
this road, the analyst searches for correspondences between concepts and the
situation described in the case. For example, when our manager is confronted
with deciding between performance and corporate values, he reframes the
issue as “performance in the context of values.” Yet, despite the reframing, he
is influenced by financial performance data during the last quarter (availabil-
ity bias). The next step along the avenue consists of inducing relationships
between concepts and actions. Here, the analyst goes beyond describing the
correspondence by showing how the bias concepts influenced the decision
made by the manager. For example, while the reframing took both criteria
into account, the final decision relied more on the recent financial informa-
tion. Approaching the end of the avenue, the analyst discovers commonalities
between the concepts and situation. This consists of an appraisal of the appli-
cation: How did these concepts contribute to understanding the dilemma
described in the case? Illustrated by these steps are the ideas of concept
retrieval, correspondence search, inducing relationships, and abstracting
commonalities. This is similar to the CBR process discussed above (e.g.,
Holyoak & Thagard, 1997). Together, they capture the process of inductive
thinking which may be a basis for a theory of case analysis (see Tanenbaum
et al. [2006[AQ: 1]], for a promising model of inductive learning and reason-
ing from a Bayesian perspective).
The design versus lecture-only comparisons provide evidence for the
broad relevance of designer learning. Strong gains for design on the bias
concepts are comparable to the earlier gains demonstrated with three negotia-
tion concepts: The sizes of the differences obtained in the two experiments
are similar (see Druckman & Ebner, 2008). The case analysis versus lecture-
only comparisons also showed short-term learning gains for the “treatment.”
This suggests that the case analysis exercise bolsters conceptual understand-
ing obtained from lectures. The inconsistent results from the earlier studies
on case–lecture comparisons suggest the possibility that the gains obtained in
this experiment may be specific to the conceptual domain investigated.
However, the case–lecture comparisons made in Experiment II are weaker
than those obtained following a design exercise used in Experiment I: Design
provides more value added (compared with lecture-only) than case analysis
in the context of learning the bias concepts. These findings augur well for the
twin goals of developing theories of synthetic and inductive learning.
The findings also contribute to the debate between Kirschner et al. (2006)
and Schmidt et al. (2007) on amount and type of guidance. The design and
Druckman and Ebner 23

case analysis exercises are compatible with the flexible adaptation approach
of PBL. Working together in dyads or small groups, the students in our exper-
iments engaged in a flexible application of the bias concepts by either design-
ing a script or analyzing a case. Regarded perhaps as an intermediate amount
of guidance, similar to assisted DL, these exercises were shown to augment
the learning obtained in a classroom lecture. They support PBL approaches to
learning but differentiate between two modalities with implications for teach-
ing: Although both modes enhance effective learning, design is likely to have
stronger impacts. Yet to be considered are possibilities for combining the
modes.
The three modes of learning compared in this study were treated as being
mutually exclusive. This is useful for experimental analyses of relative effects
on concept understanding and the findings summarized above provide helpful
insights. However, there is also reason to believe that the modes can be used
in combination. Knowledge gleaned from experiments that compare various
combinations of the modes would contribute to further theory development.
But combining the modes has practical consequences for the classroom as
well. Some leads on how to combine them are found in the process-learning
literature. Building on Whitehead’s (see Allan, 2012) idea of stages of learn-
ing, we may construe a sequence that goes from lecture-based abstract learn-
ing (as in our lecture condition) to applying the concepts in DL exercises, first
perhaps as design (invention) and then as case analysis (induction). Building
on Allan’s (2012) idea of cyclical learning or feedback loops, we can explore
the way that the three modes affect each other by exploring an interplay among
the cognitive processes of abstraction, synthesis, and induction as these are
experienced in the classroom. Looking forward, the nexus between theory and
practice is the next challenge for DL researchers.

Appendix
Please insert your assigned number:
The following six questions ask about the extent to which the exercise
contributed to your learning of each of the concepts. Please circle a phrase on
reach scale that best captures your view of learning.

1. To what extent did the exercise add to your understanding of the con-
cept of “overconfidence” presented in the lecture?

Added Added pretty Added Added only Did not Did not add
considerably much somewhat a little add much anything at all
24 Journal of Management Education 00(0)

2. To what extent did the exercise add to your understanding of the con-
cept of “framing” presented in the lecture?

Added Added pretty Added Added only Did not Did not add
considerably much somewhat a little add much anything at all

3. To what extent did the exercise add to your understanding of the con-
cept of “anchoring” presented in the lecture?

Added Added pretty Added Added only Did not Did not add
considerably much somewhat a little add much anything at all

4. To what extent did the exercise add to your understanding of the con-
cept of “availability bias” presented in the lecture?

Added Added pretty Added Added only Did not Did not add
considerably much somewhat a little add much anything at all

5. To what extent did the exercise help you understand the relationships
among the four concepts?

Added Added pretty Added Added only Did not add Did not add
considerably much somewhat a little much anything at all

6. Do you think that you would have learned more, less or the same by
only hearing the lecture on these bias concepts?

Would have learned more About the same Would have learned less

The following questions give you an opportunity to demonstrate your


understanding of each of the four bias concepts. The first question on each of
the biases is definitional, the second is an application in the workplace. Feel
free to use more space than provided if needed.

1. What is meant by “overconfidence bias?”


2. How would you identify this bias in your work as a manager?
3. What is meant by “framing?”
4. How would you identify this bias in your work as a manager?
5. What is meant by “anchoring?”
6. How would you identify this bias in your work as a manager?
Druckman and Ebner 25

7. What is meant by “availability bias?”


8. How would you identify this bias in your work as a manager?

Thanks. If you have any concerns about the conduct of this research project,
you can contact the Executive Officer at Melbourne Business School.

Acknowledgments
Special thanks go to Jill Klein at the Melbourne Business School for having her stu-
dents participate in these experiments. Jill also deserves appreciation for serving as
the lecturer for the 2014 and 2015 classes on cognitive bias in decision making. Our
thanks go out, as well, to the article’s anonymous reviewers and the Journal of
Management Education editors, for their very helpful comments.

Declaration of Conflicting Interests


The author(s) declared no potential conflicts of interest with respect to the research,
authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or publi-
cation of this article.

Notes
1. The simulation role-play literature is replete with claims, and some evidence,
about its learning advantages (see Druckman & Ebner, 2013, for a review).
Considerably less attention has been given to simulation design, particularly as an
alternative to role-playing. Thus, the comparison filled a lacuna in this literature.
2. Another helpful resource for getting a broad view of the research in this area is
the annotated bibliography maintained by the National Center for Case Study
Teaching in Science, available online at http://sciencecases.lib.buffalo.edu/cs/
pdfs/Research_Articles.pdf

References
Alfieri, L., Brooks, P. J., Aldrich, N. J., & Tenenbaum, H. R. (2011). Does discovery-
based instruction enhance learning? Journal of Educational Psychology, 103(1),
1-18.
Allan, G. (2012). Modes of learning: Whitehead’s metaphysics and the stages of edu-
cation. Albany: SUNY Press.
Bain, K. (2004). What the best college teachers do. Cambridge, MA: Harvard
University Press.
Berk, L. E., & Winsler, A. (1995). Scaffolding children’s learning: Vygotsky and early
childhood education. Washington, DC: National Association for the Education of
Young Children.
26 Journal of Management Education 00(0)

Bevan, D., & Kipka, C. (2012). Experiential learning and management education.
Journal of Management Development, 31, 193-197.
Bligh, D. (1998). What’s the use of lectures? Exeter, England: Intellect.
Bruner, J. S. (1961). The act of discovery. Harvard Educational Review, 31, 21-32.
Burgoyne, J., & Cooper, C. L. (1976). Research on teaching methods in manage-
ment education: Bibliographical examination of the state of the art. Management
International Review, 16, 95-102.
Carrero, E., Gomar, C., Penzo, W., & Rull, M. (2007). Comparison between lecture-
based approach and case/problem-based learning discussion for teaching pre-
anaesthetic assessment. European Journal of Anaesthesiology, 24, 1008-1015.
Chi, M. T. H. (2009). Active-constructive-interactive: A conceptual framework for
differentiating learning activities. Topics in Cognitive Science, 1, 73-105.
Clark, R. E. (2009). How much and what type of guidance is optimal for learning from
instruction? In S. Tobias & T. M. Duffy (Eds.), Constructivist theory applied to
instruction: Success or failure? (pp. 158-183). New York, NY: Taylor & Francis.
Crookall, D. (1995). A guide to the literature on simulation/gaming. In D. Crookall &
K. Arai (Eds.), Simulation and gaming across disciplines and cultures (pp. 151-
177). Thousand Oaks, CA: Sage.
Dempsey, J. V. (2010). Elemental learning and the pyramid of fidelity. In R. Van Eck
(Ed.), Gaming and cognition: Theories and perspectives from the learning sci-
ences (pp. 82-84). Hershey, PA: IGI Global.
Druckman, D. (2005). Doing research. Thousand Oaks, CA: Sage.
Druckman, D., & Ebner, N. (2008). Onstage, or behind the scenes? Relative learning
benefits of simulation role-play and design. Simulation & Gaming, 39, 465-497.
Druckman, D., & Ebner, N. (2013). Games, claims, and new frames: Rethinking the
use of simulation in negotiation education. Negotiation Journal, 29, 61-93.
Duke, R. D. (1974). Gaming: The future’s language. New York, NY: Halsted Press.
Ellington, H., Fowlie, J., & Gordon, M. (1998). Using games and simulations in the
classroom. London, England: Routledge.
Elsner, E. W. (1991). The enlightened eye: Qualitative inquiry and the enhancement
of educational practice. New York, NY: Macmillan.
Faria, A. J., Hutchinson, D., Wellington, W. J., & Gold, S. (2009). Developments
in business gaming: A review of the past 40 years. Simulation & Gaming, 40,
464-487.
Fletcher, J. D. (2009). From behaviorism to constructivism. In S. Tobias & T. M.
Duffy (Eds.), Constructivist theory applied to instruction: Success or failure?
(pp. 242-263). New York, NY: Taylor & Francis.
Freeman, S., Eddy, S. L., McDonough, M., Smith, M. K., Okoroafer, N., Jordt, H.,
& Wenderoth, M. P. (2014). Active learning increases student performance in
science, engineering, and mathematics. Proceedings of the National Academy of
Sciences of the United States of America, 111, 8410-8415.
Gamble, E. N., & Jelley, R. B. (2014). The case for competition: Learning about
evidence-based management through case competition. Academy of Management
Learning & Education, 13, 433-445.
Druckman and Ebner 27

Gentner, D., & Holyoak, K. J. (1997). Reasoning and learning by analogy. American
Psychologist, 52, 32-34.
Greenblat, C. S. (1981). Teaching with simulation games: A review of claims and
evidence. In C. S. Greenblat & R. D. Duke (Eds.), Principles and practices of
gaming-simulation (pp. 139-153). Beverly Hills, CA: Sage.
Holyoak, K. J., & Thagard, P. (1997). The analogical mind. American Psychologist,
52, 35-44.
Kagan, D. M. (1993). Contents for the use of classroom cases. American Educational
Research Journal, 30, 703-723.
Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why minimal guidance dur-
ing instruction does not work: An analysis of the failure of constructivist, dis-
covery, problem-based, experiential, and inquiry-based teaching. Educational
Psychologist, 41, 75-86.
Knowles, M. S. (1984). Andragogy in action. San Francisco, CA: Jossey-Bass.
Knowles, M. S., Holton, E. F., & Swanson., R. A. (2011). The adult learner. New
York, NY: Routledge.
Kolodner, J. L. (1997). Educational implications of analogy: A view from case-based
reasoning. American Psychologist, 52, 57-66.
Kunselman, J. C., & Johnson, K. A. (2004). Using the case method to facilitate learn-
ing. College Teaching, 52, 87-92.
Mayer, R. E. (2004). Should there be a three-strikes rule against pure discovery
learning? The case for guided methods of instruction. American Psychologist,
59, 14-19.
McCarthy, P. R., & McCarthy, H. M. (2006). When case studies are not enough:
Integrating experiential learning into business curricula. Journal of Education for
Business, 81, 201-214.
Merriam, S. B., Caffarella, R. S., & Baumgartner, L. S. (2006). Learning in adult-
hood: A comprehensive guide. San Francisco, CA: Jossey-Bass.
Moore, D. A., & Healy, P. J. (2008). The trouble with overconfidence. Psychological
Review, 115, 502-517.
Moore, S. (1999). Cases vs. lectures: A comparison of learning outcomes in under-
graduate principles of finance. Journal of Financial Education, 25, 37-51.
Pea, R. D. (2004). The social and technological dimensions of scaffolding and related
theoretical concepts of learning, education, and human activity. Journal of the
Learning Sciences, 13, 423-451.
Puntambekar, S., & Hubscher, R. (2005). Tools for scaffolding students in complex
learning environments. Educational Psychologist, 40, 1-12.
Roehler, L. R., & Cantlon, D. J. (1997). Scaffolding: A powerful tool in social con-
structivist classrooms. In K. Hogan & M. Pressley (Eds.), Scaffolding student
learning: Instructional approaches and issues (pp. 6-42). Cambridge, MA:
Brookline Books.
Rosenshine, B. (2009). The empirical support for direct instruction. In S. Tobias & T.
M. Duffy (Eds.), Constructivist theory applied to instruction: Success or failure?
(pp. 201-220). New York, NY: Taylor & Francis.
28 Journal of Management Education 00(0)

Salas, E., Wildman, J. L., & Piccolo, R. F. (2009). Using simulation-based train-
ing to enhance management education. Academy of Management Learning &
Education, 8, 559-573.
Schmidt, H. G., Loyens, S. M. M., van Gog, T., & Paas, F. (2007). Problem-based
learning is compatible with human cognitive architecture: Commentary on
Kirschner, Sweller, and Clark (2006). Educational Psychologist, 42, 91-97.
Taylor, S. E. (1982). The availability bias in social perception and interaction. In
D. Kahneman, P. Slovic, & A. Tversky (Eds.), Judgment under uncertainty:
Heuristics and biases (pp. 190-200). Cambridge, England: Cambridge University
Press.
Thistlethwaite, J. E., Davies, D., Ekeocha, S., Kidd, J. M., MacDougall, C., Matthews,
P., . . .Clay, D. (2012). The effectiveness of case-based learning in health profes-
sional education. Medical Teacher, 34, 421-444.
Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and
biases. Science, 185, 1124-1131.
Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology
of choice. Science, 211, 453-458.
Whitehead, A. N. (1929). The aims of education and other essays. New York, NY:
Free Press.
Wood, D., Bruner, J. S., & Ross, G. (1976). The role of tutoring in problem solving.
Journal of Child Psychology and Psychiatry, 17, 89-100.

View publication stats

You might also like