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

net/publication/289062010

Compound Problem Solving: Insights from the Workplace for Engineering

Education

Article  in  Journal of Professional Issues in Engineering Education and Practice · October 2011

CITATIONS READS

9 571

2 authors, including:

Johannes Strobel
University of Texas at El Paso
183 PUBLICATIONS   3,572 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Simulations and Modeling View project

Creativity in Engineering View project

All content following this page was uploaded by Johannes Strobel on 12 May 2016.

The user has requested enhancement of the downloaded file.

 Compound Problem Solving: Insights from
the Workplace for Engineering Education
Johannes Strobel1 and Rui (Celia) Pan2

Abstract: For engineering educators who incorporate real-world problems into their teaching, it is essential to understand real-world prob-
lem solving and the nature of problems for better design of the instruction. Prior research provided evidence that real-world problems involve
many stakeholders, are complex, and have multiple solutions and solution paths. The purpose of this paper is to provide additional insights
particularly into the structure/class of workplace problems and associated models of expertise. This paper describes the findings of a study
that employed a two-step process: (1) a single-case study of a steel engineer generating a model of compound problem solving and (2) a multi-
case comparison of 90 problem-solving narratives of other engineers to the single case. The study is located in a U.S. context. Results indicate
that real-world problems are intertwined problems (compound problems) and that transitions from one problem type to another within a
compound problem are a unique class of problems themselves. These transition problems have properties that are not represented in other
problem types and provide insights into expertise and expertise development in problem solving. DOI: 10.1061/(ASCE)EI.1943-5541
.0000047. © 2011 American Society of Civil Engineers.
CE Database subject headings: Engineering education; Engineering profession; Engineers; Management methods; Research; Publications.
Author keywords: Engineering education; Engineering profession; Engineers; Management methods; Research; Publications.

Introduction resources, optimum, and uncertainty. Jonassen et al. (2006)

For years, reports have validated the importance of problem solving
in the workplace. For instance, the report by Sheppard et al. (2008)
stated that problem solving is an essential thinking skill for engi-
neers. Professionals in any field are hired, retained, and rewarded
for their abilities to solve workplace problems. For engineering ed-
ucation, this means a challenge to prepare students to solve real-
world problems.
Several instructional and educational strategies heavily utilize
real problems, such as problem-based learning (PBL), project-
based teaching, or service learning (Savery 2006). These ap-
proaches share the aspiration to integrate real-world workplace
problems into the curriculum and face the same challenges of
staying abreast with new and changing roles of engineers in the
workplace. Understanding problems and problem solving in
the workplace is essential to better design problems for class-
rooms, better design support structures for students engaging in
problem-oriented instruction, and research the effectiveness on
students’ performance and conceptual development.
Several models exist that address the categorization and unique
attributes of real-world problems. Koen (2003) proposed four
characteristics of an engineering workplace problem: change,
1
Ph.D., Director, INSPIRE, Institute for P-12 Engineering Research
and Learning, 315 N. Grant St., West Lafayette, IN 47907; and Assistant
Professor, Engineering Education & Educational Technology, Purdue
Univ., Neil Armstrong Hall of Engineering, 701 W. Stadium Blvd., West
Lafayette, IN 47907 (corresponding author). E-mail: jstrobel@purdue.edu
2
Ph.D. Student, School of Engineering Education, Purdue Univ., Neil
Armstrong Hall of Engineering, 701 W. Stadium Blvd., West Lafayette,
IN 47907.
Note. This manuscript was submitted on November 17, 2009; approved
on July 29, 2010; published online on August 31, 2010. Discussion period
open until March 1, 2012; separate discussions must be submitted for in-
dividual papers. This paper is part of the Journal of Professional Issues in
Engineering Education & Practice, Vol. 137, No. 4, October 1, 2011.
©ASCE, ISSN 1052-3928/2011/4-215–222/$25.00.
provided evidence for distinct features of workplace engineering
problems, for example, that workplace problems can be solved
in many ways, are ill-structured and complex, and success is rarely
measured solely by engineering standards. One of the most com-
prehensive models is Jonassen’s design theory of problem solving
in which he described 11 different problem types mapped on a
four-dimensional scale (Jonassen 2000).
Although these problem types are helpful to classify problems,
real-world problems, as proposed by Jonassen, are more likely to be
metaproblems or compound problems containing a variety of dif-
ferent problem types. Although there is a growing body of literature
on researching differences in solving the variety of different
problem types (Uribe et al. 2003), little research provides evidence
for and understanding of compound problems. In addition, as
Trevelyan and Tilli (2007, p. 300) pointed out, “there are only a
few accounts of the people that do engineering work and what that
work actually involves.”
The purpose of this research study is to provide insights on these
gaps in the literature, and it sets out to accomplish the following
high-level agenda:
a. Provide evidence for the existence of compound problems and
their attributes,
b. Research the problem-solving processes of engineers while en-
gaged in compound problem solving in the workplace, and
c. Recognize a new class of skills (compound problem solving),
which extends existing models of expertise and expertise devel-
opment in problem solving.
Theoretical Foundation
Problem-based and project-based learning curricula have been
implemented in numerous engineering programs, for example,
biomedical engineering (LaPlaca et al. 2001) and chemical engi-
neering (Cline and Powers 1997). In his framework on designing
problems for a problem-based learning curriculum, Hung (2006)

JOURNAL OF PROFESSIONAL ISSUES IN ENGINEERING EDUCATION & PRACTICE © ASCE / OCTOBER 2011 / 215

Downloaded 10 Nov 2011 to 128.210.126.199. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
pointed out that it is crucial for the curriculum design to align
systematically (a) the nature of the problem, (b) the nature of
the skills involved, (c) the content material, (d) the goals to achieve,
and (e) the strategies to be employed. In addition, curricula face a
continuous challenge to maintain their base of problems authentic
and informed by the everyday practice of practitioners. This—
together with a systematic process of identifying attributes of work-
place problems—entails the challenge to understand the underlying
mechanism and cognitive and sociocognitive processes, which are
at work when people solve problems.
Conceptualizations of Problems and Problem Solving
Common types of problems utilized in engineering education are
well-structured problems. Well-structured problems have the
parameters specified in the problem statement; possess knowable,
correct solutions that are determined by preferred solution paths;
and apply a limited number of regular rules and principles that
are organized in a predictive and prescriptive arrangement (Voss
and Post 1988). A specific form of a well-structured problem is
the story (word) problem. When learning to solve story problems
in engineering, students learn to translate relationships about un-
knowns into equations, solve the equations to find the value of
the unknowns, and check the values found to see if they satisfy
the original problem (Rich 1960). This linear process implies that
solving problems is a procedure to be memorized, practiced, and
habituated, which emphasizes obtaining the answer over obtaining
the meaning (Wilson 1993).
Educators have historically assumed that a general (and generic)
problem-solving model is sufficient and problem-solving skills
acquired through working on simple and well-defined problems
would easily transfer into more complex and ill-defined problem-
solving tasks (Salvucci and Anderson 2001). Not surprisingly, most
instructional materials like textbooks utilize well-structured prob-
lems because they were considered to sufficiently prepare learners
for more ill-structured and complex problem-solving tasks.
Assumptions on transferring well-defined problem solving to
ill-defined problem solving were challenged by a wide variety
of researchers (for a summary, see Mayer 1998) arguing that prob-
lems vary by nature, context, constraints, and problem-solver char-
acteristics (Hong et al. 2003). That is, learning to solve story
problems in schools does not enable graduates to solve complex
and ill-structured problems. Ill-structured problems have vaguely
defined or unclear goals and unstated constraints; possess multiple
solutions and solution paths or no consensual agreement on the
appropriate solution; involve multiple criteria for evaluating solu-
tions; possess no explicit means for determining appropriate actions
or relationships between concepts, rules, and principles to be used;
and require learners to make judgments, express personal opinions
and choices about the nature of the problem and proposed solution
strategies, and defend them (Jonassen 1997).
Additional conceptualizations on problem solving emerged,
which include several dimensions expanding the well-structured/
ill-structured dimension. Mayer and Wittrock (1996) referred to
ill- and well-defined and routine–nonroutine problems. By using
a multidimensional model, Jonassen’s (1997, 2000) framework
provided a taxonomy of problems with classes of attributes. The
dimensions are ill and well structured, complex and simple, situated
and abstract, context dependent and context independent, and static
and dynamic. These five dimensions can be utilized to specify
attributes of 11 problem variations ranging from logic problems
over troubleshooting and design to dilemma problems, to name
a few (see Fig. 1 for an illustration mapping different problem types
on the five different dimensions).
216 / JOURNAL OF PROFESSIONAL ISSUES IN ENGINEERING EDUCATION & PRACTICE © ASCE / OCTOBER 2011
Downloaded 10 Nov 2011 to 128.210.126.199. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
Fig. 1. Problem types mapped on five dimensions, on the basis of
Jonassen (2000)
Jonassen proposed that the individual types of problems are ar-
chetypes and rarely occur pure, and the distinction rather serves an
analytical purpose. He further postulated that real-world problems
consist of a combination of many problem variations, rendering
them to be metaproblems (2000, p. 81). Yet, the attributes of these
metaproblems and the problem-solving skills to solve metapro-
blems remain unclear.
Problem-Solving Processes and Expertise
Research on classifying problem types needs to be augmented with
an understanding of the cognitive processes of problem solvers and
a model of how an expertise in problem-solving manifests. Existing
research in workplace problem solving was conducted through a
close examination of an expert’s performance in solving real-world
problems.
Expertise can be defined as “the possession of a large body of
knowledge and procedural skills” (Chi et al. 1982). In recent
years, research in expertise has received increasing attention, most
of which involves studies of experts and comparison of experts’ and
novices’ performance.
In general, experts perform better than novices in a number
of respects. By reviewing present studies in expertise, several
common features of expertise can be identified across different
areas, such as experts’ ability to recognize similarities among sit-
uations (Batra and Davis 1992; Crismond 2001), the ability to form
abstract conceptualizations and decompose problems explicitly
(Cross 2004), and experts’ skill at such things as perceiving large
meaningful patterns, encoding new information quickly, adapting
decision strategies to changing task conditions, and generating a
holistic understanding of a problem (Cellier et al. 1997; Farrington-
Darby and Wilson 2006). In essence, Cross (2004) described ex-
perts to be “ill-behaved” problem solvers, which means they can
adjust their skills and recall experiences to find unique solutions
to similar and new problems while straying from existing paths.
It takes a long period of time to become a real expert. As
Dreyfus and Dreyfus (1986) identified, novices go through
progressive stages from novice, advanced beginner, competent,
and proficient before finally becoming an expert. Thus, the trans-
formation from novice to expert requires a great length of time; as
summarized by Ericsson et al. (1993), the time required for a novice
to attain exceptional performance may be over 10 years.
However, the learning of expertise may be accelerated. Ericsson
et al. (1993) claimed that deliberate practice is a key factor in
expertise development. One example of this practice is the
investigate-and-redesign (I&R) task proposed by Crismond (2001).
Although current research suggests that expertise is largely do-
main specific, it is unclear if attributes of expertise apply unequivo-
cally to compound problem solving. Similarly, it is unclear which
role the repertoire of strategies plays for solving different problems
and how experts activate and switch between strategies. Virtually
no studies reveal expertise in compound problem solving, and
because different types of problems intertwine with one another
within compound problems, experts moving between various kinds
of problems may utilize unique strategies. Consequently, further
research in expertise in compound problem solving is needed to
investigate whether existing knowledge in expertise could explain
performance in compound problem solving.
Purpose of the Study
The purpose of this study is to provide insights on these gaps in the
literature by answering the following specific research questions:
1. How do different problem variations interplay within so-called
compound problems?
2. When engaged in a compound problem, how does a problem
solver switch between the different problems and from one
problem-solving mode required by one problem type into
another mode required by another problem type?
3. How can the different modes of compound problem solving
contribute to/extend knowledge on expertise in problem
solving?
Methodological Framework
As methodological framework, the writers employed a modified
analytic induction (Bogdan and Biklen 1992) process, a qualitative
research methodology that systematically uses a set of five steps
to develop an inductively derived grounded theory (Strauss and
Corbin 1990). The five steps include: (1) define and explain the
particular phenomenon; (2) examine the definition and explanation
through the data collected; (3) modify the definition and explana-
tion, if new data make it necessary; (4) deliberately attempt to find
data that may not fit the explanation/definition; and (5) repeat until
the definition/explanation embraces all data and accounts for all
discrepancies. Modified analytic induction provided a methodol-
ogy for identifying themes and categories within engineering sto-
ries told by practicing engineers. In the analytical induction
approach, data built the basis for further descriptions and interpre-
tations, but as the term induction indicates, the methodology did
not employ a theoretical empiricism, but rather was informed by
prior research. This research methodology seemed very appropriate
for answering fundamental research questions about the nature of
workplace problem solving.
To operationalize this research, the following two steps were
taken: (1) a single-case study investigated problem-solving proc-
esses of one engineer in two distinct problem-solving activities,
a design task and a trouble-shooting task, which generated a model
of compound problem solving and cumulated into a rubric to assess
problem solving; (2) the model and rubric generated in the first step
was applied to 90 previously documented cases of problem solving
by engineers to compare and contrast findings from the first step
across a larger sample.
Step 1: Single-Case Study
The research in the first step focused primarily on the particular
types of problems encountered and the processes of problem solv-
ing employed by an engineer during a troubleshooting project and a
design project in the context of steel mills for the production of
tubes and pipes.
JOURNAL OF PROFESSIONAL ISSUES IN ENGINEERING EDUCATION & PRACTICE © ASCE / OCTOBER 2011 / 217
Downloaded 10 Nov 2011 to 128.210.126.199. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
Participant
The research participant in the single-case study was a 65-year-old
male engineer who holds a B.Sc. in Electrical and Mechanical
Engineering. After working as an electrical engineer on radar sys-
tems in the 1960s, he switched in the 1970s to the steel industry and
continues to work as an independent general contractor and con-
sultant. His clients range from small steel mill installations to large
multinational corporations producing tubes and pipes as their core
business. He is considered an expert in turnkey solutions for new
steel mills, redesigning and extending existing steel mills, and trou-
ble shooting systems that have severe performance problems.
Data Sources
For the single-case study, the following data sources were utilized:
(a) design drawings, specification with documents, and contract
information; (b) 10 formal interviews and numerous informal inter-
views; and (c) informal observations during design and trouble-
shooting sessions.
Data Collection
For the first step, data were collected over the period of a year and
drew from two related but different projects. In the first project, the
installation of a large steel mill was nearly complete, though fine-
tuning, troubleshooting, and assessment questions were in the fore-
ground. The second project started within the second month of the
year and was focused on early-stage constraints, design issues,
and goals.
During the first meetings, the researcher asked the engineer to
describe in as much detail as possible what he knew about the two
projects, including goals, constraints, foreseeable problems, and
possible solutions. In the remainder of the interviews, the engineer
was asked to explain and walk the researcher through different
steps of the problem and to describe changes in his knowledge
about the problem, solution strategies, and solutions. The formal
interviews were held irregularly with the time between interviews
ranging from 3 weeks to 2 months; the interviews were each 2 to
3 h in length.
During the time of the observations, the engineer was drawing,
reviewing drawings, making estimates, performing calculations,
contacting vendors and his clients, and inspecting equipment.
During the observation periods, the engineer was prompted to
describe the processes and engage in metacognitive reasoning,
especially in building decision and constraint trees (Johnson and
Johnson 1991). The data collection in this phase also included
his drawings and estimates; these were later annotated. There were
three observation periods totaling approximately 5 h.
Data Analysis
For the qualitative analysis of all the available data, the interviews
were transcribed. The graphical material was annotated and con-
nected with the interview transcripts if pertinent for the interview
topics. The observation notes were sorted and initial connections to
the other forms of data were documented. To make sense of the
data, all the material was first read to develop initial coding
schemes until emerging categories were saturated (Strauss and
Corbin 1990), meaning no new dimensions appeared from the data.
On the basis of the coding scheme, the data were then analyzed and
coded in depth. The coding rubric was designed to establish differ-
ent problem attributes, the emergence of different problems within
problems, and the change over time in conceptualizing the problem,
determining the solution strategies, and evaluating the solution.
The different problems were treated as multiple cases. Bertaux
noted the importance of several cases “which together help the
investigator understand what is called the object” (1981, p. 9).
In this study, the objects were the threefold research questions as
previously laid out. By utilizing a multicase look at the different
problems, several problems and problem-solving strategies were
compared and contrasted with one another and with the existing
conceptualizations of Jonassen (1997, 2000).
With a modified analytic induction, the multicase lenses
within a single-case study design were expected to produce addi-
tional insights into the three research questions and facilitate the
“bracketing” (Bogdan and Biklen 1992) of prior conceptualiza-
tions. Strauss and Corbin defined this approach as “a qualitative
research method that uses a systematic set of procedures to develop
an inductively derived grounded theory about a phenomenon”
(1990, p. 24).
Collecting and coding the material constituted step one of the
constant comparative analysis. Codes are abbreviations or symbols
applied to a segment of words to facilitate sorting and clustering
word segments relating to a particular topic or question (Bogdan
and Biklen 1992). By using the guiding questions, the first
writer developed categories of information (open coding). In the
open coding phase, the first writer examined the textual and visual
information (transcripts and drawings) for salient categories of in-
formation supported by the text. Further, the theories were intercon-
nected (axial coding), a story was built that connects the categories
(selective coding), and a discursive set of theoretical propositions
(with a single category as the central phenomenon in the center)
was created to fulfill the last three steps of the comparative analysis
(Strauss and Corbin 1990).
In the process of analysis, a qualitative research tool (Qualrus)
was utilized to manage the data. It allowed coding of texts and
multimedia materials (including the drawings) and provided a
variety of tools in the development of theory out of data, including
a concept-mapping tool.
Results and Discussion—Model of Compound Problem
Solving and Rubric
The findings of the single-case analysis can be structured into
different areas: (a) problem-solving processes, especially within
compound problems; (b) extensions of design theory of problem
solving; and (c) reflections and extensions of expertise models
of problem solving.
The two projects could be broadly categorized as (1) a trouble-
shooting problem in the process of setting up a steel mill and adjust-
ing the steel mill configurations for different needs and (2) a design
process project. However, several aspects emerged that put the pre-
liminary categorization of the design problem in question, as is
shown by the following example description: The engineer in-
stalled a new production line that should produce 200 steel tubes
in 20 cycles at 2 min/cycle. On the one hand, the design problem
in this particular case is very well structured—all specifics of the
current state (need, existing production capacity) and the goal state
(specifics of the solution) are known. The process of solving the
problem, on the other hand, looks very different: there are multiple
tools, processes, considerations, and decisions to find the most
cost-effective and efficient solution. In summary, the problem is
very well structured whereas the process of solving it is very ill
structured.
For further exploration of the process of solving the problem,
other differentiations have to be taken into consideration. The
solving of this design problem contains several complex and ill-
structured decision processes (weighing of options and forecasting
the impact of decisions on a wide variety of variables). The process
also contains procedural and well-structured components, like the
measuring of values and the installation of components. In existing
classifications, this problem would have been conceptualized
as ill-structured and complex. The differentiation between the
218 / JOURNAL OF PROFESSIONAL ISSUES IN ENGINEERING EDUCATION & PRACTICE © ASCE / OCTOBER 2011
Downloaded 10 Nov 2011 to 128.210.126.199. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
well-structured problem statement (specifications), the ill-
structured process of finding an acceptable solution, and the well-
structured but complex procedural steps to change specifications or
measure the current state of the problem questions this classifica-
tion. The study confirmed previous conceptualizations that the
project, even categorized as a design problem, contained many dif-
ferent other problem variations required to solve the problem in its
entirety (Zeng 2002, 2004).
Another example from the troubleshooting problem at a steel
mill (the first project) will further demonstrate this point. In this
project, a steel mill has a complex performance problem. There
is a 5 s delay when one cycle ends and the next starts. With thou-
sands of cycles per year, the economic implications of this problem
are considerable for the company. The current state is unclear, the
goal state is very clear (the machine should run again under similar
specifications as before without the 5 s delay); the process is ill
structured (since the current state is not clear, it is not clear what
the error is, and therefore not clear what solution path to take). The
analysis of the error is at the core of this particular problem. If we
look deeper into the process of troubleshooting, we find that it is
again a mix between well-structured and ill-structured problems.
The process of collecting data is very well structured (well-
described procedures of measuring malfunctions), the interpreta-
tion of the data and the decisions about which direction to go from
the findings is very ill-structured. The actual solving of the problem
is again very well-structured (well-defined procedures of how to
change parts or change actual specific settings).
The question of classification seems at first purely academic.
There are, however, huge implications in teaching students
problem-solving skills, designing and developing problems, and
providing support during problem-solving activities. The project
sheds light on problem solving in the context of an expert engineer.
Expert problem solving is often conceptualized as seeing the bigger
picture (Chaffin et al. 2003; Azevedo 1998; Sutherland 2002)
and rather conceptually than procedurally understanding a problem
(Wijekumar 2001; Kushniruk 2000; Ferrell-Swann 1999). This
study challenges the strong separation between conceptual and
procedural knowledge in describing problem solving, as similarly
argued by Barnett and Koslowski (2002).
As seen in the preceding findings, the design theory does not
make a distinction between the process and the solution. It sub-
sumes the process under the structuredness of the problem. The
research conducted in this study would suggest that there is a need
to make a distinction between the structure of the problem and the
structure of the process of solving the problem. The previously
mentioned example descriptions show that problem solving of
so-called ill-structured problems becomes more complex because
it seems to be necessary to take internal transitions within a prob-
lem (from well-structured to ill-structured) into account.
The research findings of the first study suggest breaking down
each problem type, separating (a) the process to identify the actual
state, (b) the nature of the actual state, (c) the process to identify the
goal state, (d) the nature of the goal state, (e) the process of problem
solving, (f) the determination of success, and (g) the measurement
of success. The dimensions to classify problems as introduced by
Jonassen (2004), structuredness, complex/simple, static/dynamic,
context (inter-) dependent, need then to be applied to the different
parts of the broken-down problem. The breakdown would support a
clearer classification of the problem types and would provide a bet-
ter analytical tool to determine switches between problem types
(the emphasis of the second step in the research study, which will
be subsequently described).
As described in the literature review, expertise can be defined as
“the possession of a large body of knowledge and procedural skills”
(Chi et al. 1982). As expertise consists of general transferable and
domain-specific skills, the research presented here similarly dem-
onstrates that expertise is a compound mix of knowledge and pro-
cedural skills. Our study additionally seems to indicate that the
management of different skill sets and knowledge within a project
and the deliberate shift between one area of expertise to another are
different sets of skills and require different knowledge from the per-
forming expert. The research presented here seems to further indi-
cate that expertise does not unequivocally apply to compound
problem solving and the engineer in this step of the research study
utilizes unique problem-solving strategies to move between various
kinds of problems within a compound problem. The research par-
ticipant is able to work with the well-structured components of the
problem, has the experience and the skills to put preliminary find-
ings in context, and is able to decide in an ill-structured situation
where to go next. In summary, he seems to be able to switch seam-
lessly between the well structured and ill structured and between
the simple and complex components of the problem and is able
to match his skills to the particular problem-solving process in
front of him, confirming earlier conceptualizations by Dicks et al.
(2005). This seamless switch seems to be a different class of
expertise.
The findings of the first step can be summarized as the
following:
1. To adequately describe workplace problems, we need to sepa-
rate the problem from the process of solving it;
2. Workplace problems are intertwined or compound, containing
differently structured subproblems that require different skills
to solve;
3. Problem-solving expertise does include both conceptual and
procedural components;
4. In addition to existing models, expertise needs to be also ex-
plained as a function of how seamless an engineer can switch
between different compound problem types; and
Fig. 2. Rubric generated from single-case study applied to 90 interviews
JOURNAL OF PROFESSIONAL ISSUES IN ENGINEERING EDUCATION & PRACTICE © ASCE / OCTOBER 2011 / 219
Downloaded 10 Nov 2011 to 128.210.126.199. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
5. The ability to shift between compound problem components is
a unique skill set that needs further exploration.
Findings from the single case were utilized to generate a rubric
(see Fig. 2). The rubric consisted of elements that were found from
the single-case study, primarily the intertwinement of different
problem types (meaning the co-occurrence of multiple problem
types within a single compound problem) and the intraproblem
transitions in the process of problem solving.
Step 2: Comparing Results with Documented
Narratives of Engineers
In the second step, the researchers compared the findings of the
single case to previously collected and analyzed stories of 90
engineers discussing their problem-solving strategies (Center for
the Study of Problem Solving 2010). Details on the creation of the
library and an initial analysis of the contained data can be found
elsewhere (Jonassen et al. 2006). The main goal of the second step
was to test the findings of the single case (the research results and
the rubric) across a larger population of cases.
Process
Data were analyzed by primarily qualitative methods, and content
and constant comparison analyses of the results of Step 1 were
applied and contrasted to the data sources of Step 3 with the aid
of the rubric created in Step 2. In addition, descriptive statistics
were applied to quantify the results.
Findings and Discussions of Step 3
As depicted in Fig. 3, the large majority of problems are intertwined
with one or more problems. Most of intertwined problem types are
different from the problem in which they are embedded (label num-
ber ending with 1). In this category, the writers classified, for ex-
ample, an emerging decision problem within a design problem and
a textbook problem that occurred during a troubleshooting prob-
lem. In the remaining intertwined problems, the same problem type

Fig. 3. Frequency of intertwined problems
occurred. Additionally, Fig. 3 can be interpreted that, when multi-
ple problems are intertwined, it is more likely that the problems are
different in nature. Further research would be necessary to test the
explanation further. This research did not separately account for
multiple instances of the same problem in sequence, demonstrated
in the following example: When the engineer described the prob-
lem as measuring success and determining solution criteria, this
was counted as one problem, even when different measurement cri-
teria and resulting different measurements were employed.
Sets of problems categorized as intertwined and different (for
example, under label B.2) defied further classification, meaning
that there was no detectable pattern/similarity between individual
sets of problems. The writers attempted to further classify and
would have ended up with 80 different combinations or sets of
different problem types, taking especially the goal, process, and
structure of the problem into consideration. The findings would in-
dicate that during the process of problem solving, new compound
problem variations emerge that seem to have unique features and
attributes that are hard to compare to one another.
Fig. 4. Frequency of types of transitions
220 / JOURNAL OF PROFESSIONAL ISSUES IN ENGINEERING EDUCATION & PRACTICE © ASCE / OCTOBER 2011
Downloaded 10 Nov 2011 to 128.210.126.199. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
Fig. 4 provides insight into one dimension in which one can
map different compound problems: the transition between different
problem-solving modes during one problem-solving activity. Only
a small number of problems in this sample of 90 interviews con-
tained a back-and-forth switch between one problem and another.
This finding is particularly surprising because earlier research
demonstrated that ill-structured problem solving often involves iter-
ation and revisiting of previous steps (Zeng 2004). A possible ex-
planation, which needs to be further tested, is that the research here
relies on narratives in the form of stories, which by nature are more
linear and not recursive. Surprisingly, the number of big deliberate
switches between different problems was rather large.
The number of big deliberate switches together with the quali-
tative data from which it is derived would also indicate that within
bigger problems, there are smaller discrete problems occurring that
follow in sequence.
An emerging finding was label D, in which the transition from
one problem to the other within the compound problem is described
as the crucial problem. Classical literature on complex and ill-
structured problem solving (especially designing) indicates that
problem solving means decomposing the complete problem into
subproblems of solvable size (see Simon 1999; Schön 1987;
Alexander 1979). Although this study corroborates the overall
description of the problem-solving process as previously described,
the findings especially represented by Fig. 4 would indicate that
because of the necessary decomposition process, the diverse nature
of the subproblems, and the overall assembly of the decomposed
problems into a whole, a new class of problems emerge: the tran-
sition problems. This provides a strong alignment between findings
from the single case and the comparisons to the 90 narratives: The
single case of the steel engineer (in Step 1) provided a new emerg-
ing definition of expertise—the switching between different prob-
lem types. As is shown in the results of the comparison of the 90
narratives (conducted in Step 3), 18 engineers (20% of the total 90)
noticed the transition between problems to the point of declaring it
a problem by itself. Further research is necessary to shed light on
the expertise level of the engineers that demonstrated more struggle
and higher success with transition problems and how to adequately
address them in the design and utilization of problems in education.
Extension of Expertise Models in Problem Solving
Defining problems as compound problems has numerous implica-
tions for defining expertise and expertise development with regard
to problem solving. Seamless transitions between different problem
types and the shift into different problem-solving modes within
a particular compound problem seem to be an important skill
set. Although research acknowledges that experts refer to a larger
reservoir of prior experience of problem-solving examples (Mayer
1998), this research indicates that past experience might have
multiple components: (a) large picture of the problem to be solved,
(b) different procedures and techniques to be activated, and
(c) mechanisms to evaluate, decide, and meaningfully switch/
transition between different techniques and thought processes.
Further, the recognition of transition or interface problems that
occur when moving from one problem-solving mode into another
during a problem-solving process are problems that are under-
researched and need further extensions, particularly when it comes
to expertise and competency development.
Implications for the Teaching of Problem Solving in
Engineering Education
This study has numerous implications for the teaching of problem
solving in engineering. Research provided here reinforces earlier
calls for more authentic problems and authentic problem solving
in engineering classrooms. It seems particularly hard to design for
the classroom problems with the same complexity and compound
nature as the problems in this study. By extending the notion of
problem solving, particularly in the context of compound problems,
it seems only consistent to revisit how problem solving is taught.
Although many instructional and curricula programs exist (e.g., for
troubleshooting, design cycle, and methods) to teach discrete forms
of problem solving, the engineering education community is well
advised to develop new teaching strategies when students encoun-
ter difficulties transitioning from a conceptual and brainstorming
activity into a more procedural and linear mode of problem solving,
for example.
Limitations of the Study and Further Research Needed
There are numerous limitations to the study. The single-case study
was situated in the steel industry and involved an engineer with an
electrical and mechanical engineering background. Whether design
and trouble-shooting problems in steel engineering are comparable
to other engineering fields needs to be further investigated. The en-
gineer in the single-case study was particularly articulate about his
thinking processes and in reflecting on the constraints of his work.
If further research wants to utilize the methodologies of this study
for different case studies, other elicitation methods need to be con-
sidered and compared to account for the fact that other research
participants might not be as able to share their reasoning processes.
This study is only a small first step in understanding the com-
plexity of the compound problem-solving enterprise. Transition
problems need to be further researched to understand the process
of problem solving and to extend further the notion of expertise in
problem solving. Additionally, the extension on the design theory
of problem solving and the accompanying model of expertise have
yet to find their way into the design of problem-solving-oriented
instruction, particularly into creation of PBL curriculum.
Conclusion
There are several conclusions that can be drawn from the data.
Workplace problems are hard to compare, especially since most
problems seem to be unique compound problems. The unique
nature of compound problems poses a particular challenge for
the design of problems and how to support the skill development
of students. Because different skills and reasoning processes are
fostered by workplace problems, instructors and the curriculum
need to account for potential difficulties in the wide spectrum of
problem-solving skills and not only in a selected few areas. This
means, for example, when students are engaged in a design activity,
they might still need support in some of the routine activities that
the design process affords. Further translational research is neces-
sary to fully integrate the research findings of this study into the
engineering curriculum.
Acknowledgments
This research was partly funded by the National Science
Foundation (NSF-0350305) (PI: David Jonassen) as part of the
activities of the Center for the Study of Problem Solving at the
University of Missouri, Columbia. Additional funding was pro-
vided by FRDP Funding, Concordia University, Montreal, Canada
(PI: Johannes Strobel) and a PRF grant at Purdue University
(PI: Johannes Strobel).
JOURNAL OF PROFESSIONAL ISSUES IN ENGINEERING EDUCATION & PRACTICE © ASCE / OCTOBER 2011 / 221
Downloaded 10 Nov 2011 to 128.210.126.199. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
References
Alexander, C. (1979). The timeless way of building, Oxford University
Press, New York.
Azevedo, R. (1998). “Expert problem solving in mammogram interpre-
tation: A visual cognitive task.” ProQuest Dissertations and Theses,
〈http://search.proquest.com/docview/304503359〉
Barnett, S. M., and Koslowski, B. (2002). “Adaptive expertise: Effects
of type of experience and the level of theoretical understanding it
generates.” Think. Reason., 8(4), 237–267.
Batra, D., and Davis, J. G. (1992). “Conceptual data modelling in database
design: Similarities and differences between expert and novice
designers.” Int. J. Man-Mach. Stud., 37(1), 83–101.
Bertaux, D., ed. (1981). Biography and society: The life history approach
in the social sciences, SAGE, Beverley Hills, CA.
Bogdan, R. C., and Biklen, S. K. (1992). Qualitative research for educa-
tion: An introduction to theories and methods, 1st Ed., Allyn & Bacon,
Boston.
Cellier, J., Eyrolle, H., and Marine, C. (1997). “Expertise in dynamic
environments.” Ergonomics, 40(1), 28–50.
Center for the Study of Problem Solving, School of Information Science
and Learning Technologies, Univ. of Missouri-Columbia. 〈http://csps
.missouri.edu/psi/cbrsearch/inventoryinterface.jsp〉, (Jul. 20, 2010).
Chaffin, R., Imreh, G., Lemieux, A. F., and Chen, C. (2003). “‘Seeing the
big picture’: Piano practice as expert problem solving.” Music Percept.,
20(4), 465–490.
Chi, M. T. H., Glaser, R., and Rees, E. (1982). “Expertise in problem solv-
ing.” Advances in the psychology of human intelligence, R. J. Sternberg,
ed., Vol. 1, Erlbaum, Hillsdale, NJ, 7–75.
Cline, M. J., and Powers, G. J. (1997). “Problem based learning via open
ended projects in Carnegie Mellon University’s chemical engineering
undergraduate laboratory.” Proc., 27th Annual Frontiers in Education
Conference, Vol. 1, IEEE, New York, 350–354.
Crismond, D. (2001). “Learning and using science ideas when doing
investigate-and-redesign tasks: A study of naive, novice, and expert
designers doing constrained and scaffolded design work.” J. Res.
Sci. Teach., 38(7), 791–820.
Cross, N. (2004). “Expertise in design: An overview.” Des. Stud.,
25(5), 427–441.
Dicks, D., Garzotto, F., Hedberg, J., and Zeng, Y. (2005). “Imagining a
science of instructional design.” Proc., World Conference on Educa-
tional Multimedia, Hypermedia and Telecommunications, P. Kommers
and G. Richards, eds., Association for the Advancement of Computing
in Education, Chesapeake, VA, 960–965.
Dreyfus, H., and Dreyfus, S. (1986). Mind over machine: The power of
human intuition and expertise in the era of the computer, Free Press,
New York.
Ericsson, K. A., Krampe, R. T., and Tesch-Römer, C. (1993). “The role of
deliberate practice in the acquisition of expert performance.” Psychol.
Rev., 100(3), 363–406.
Farrington-Darby, T., and Wilson, J. (2006). “The nature of expertise:
A review.” Appl. Ergon., 37(1), 17–32.
Ferrell-Swann, K. L. (1999). “An exploration of the components of coun-
selor expertise.” ProQuest Dissertations and Theses, 〈http://search
.proquest.com/docview/304523444〉.
Hong, N. S., Jonassen, D. H., and McGee, S. (2003). “Predictors of well-
structured and ill-structured problem solving in an astronomy simula-
tion.” J. Res. Sci. Teach., 40(1), 6–33.
Hung, W. (2006). “The 3CR model: A conceptual framework for designing
problems in PBL.” Interdiscip. J. Probl.-Based Learn., 1(1), Article 6.
Johnson, H., and Johnson, P. (1991). “Task knowledge structures: Psycho-
logical basics and integration into system design.” Acta Psychol.,
78(1–3), 3–26.
Jonassen, D. H. (1997). “Instructional design models for well-structured
and ill-structured problem-solving learning outcomes.” Educ. Technol.
Res. Dev., 45(1), 65–95.
Jonassen, D. H. (2000). “Toward a design theory of problem solving.”
Educ. Technol. Res. Dev., 48(4), 63–85.
Jonassen, D. H. (2004). Learning to solve problems: An instructional
design guide, Jossey-Bass, San Francisco.
Jonassen, D., Strobel, J., and Lee, C. B. (2006). “Everyday problem solving
in engineering: Lessons for engineering educators.” J. Eng. Edu., 95(2),
139–151.
Koen, B. V. (2003). “Some thoughts on engineering.” Chapter 1, Discus-
sion of the method: Conducting the engineer’s approach to problem
solving, Oxford University Press, New York, 7–25.
Kushniruk, A. W. (2000). “Complex decision making in intensive care: The
role of medical expertise.” Diss. Abst. Int., B, 61(6-B), 3300.
LaPlaca, M. C., Newstetter, W. C., and Yoganathan, A. P. (2001).
“Problem-based learning in biomedical engineering curricula.” Proc.,
31st Annual Frontiers in Education Conference, Vol. 2, IEEE,
New York, F3E/16–F3E/21.
Mayer, R. E. (1998). “Cognitive, metacognitive, and motivational aspects
of problem solving.” Instr. Sci., 26(1/2), 49–63.
Mayer, R. E., and Wittrock, M. C. (1996). “Problem-solving transfer.”
Handbook of educational psychology, D. Berliner and R. C. Calfee,
eds., Macmillan, New York, 47–62.
Qualrus [Computer software]. Columbia, MO, The Idea Works, Inc.
Rich, B. (1960). Schaum’s principles and problems of elementary algebra,
Schaum’s outline series, New York.
Salvucci, D. D., and Anderson, J. R. (2001). “Integrating analogical
mapping and general problem solving: The path-mapping approach.”
Cogn. Sci., 25(1), 67–110.
Savery, J. R. (2006). “Overview of problem-based learning: Definitions and
distinctions.” Interdiscip. J. Probl.-Based Learn., 1(1), 9–20.
Schön, D. A. (1987). Educating the reflective practitioner, Jossey-Bass,
San Francisco.
Sheppard, S. D., Macatangay, K., Colby, A., and Sullivan, W. S. (2008).
222 / JOURNAL OF PROFESSIONAL ISSUES IN ENGINEERING EDUCATION & PRACTICE © ASCE / OCTOBER 2011
View publication stats Downloaded 10 Nov 2011 to 128.210.126.199. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
Educating engineers: Designing for the future of the field, Jossey-Bass,
San Francisco.
Simon, H. A. (1999). The sciences of the artificial, 3rd Ed., MIT Press,
Cambridge, MA.
Strauss, A., and Corbin, J. (1990). Basics of qualitative research:
Grounded theory procedures and techniques, SAGE, Newbury Park,
CA.
Sutherland, L. (2002). “Developing problem solving expertise: The impact
of instruction in a question analysis strategy.” Learn. Instr., 12(2),
155–187.
Trevelyan, J., and Tilli, S. (2007). “Published research on engineering
work” J. Prof. Issues Eng. Educ. Pract., 133(4), 300–307.
Uribe, D., Klein, J. D., and Sullivan, H. (2003). “The effect of computer-
mediated collaborative learning on solving ill-defined problems.” Educ.
Technol. Res. Dev., 51(1), 5–19.
Voss, J. F., and Post, T. A. (1988). “On the solving of ill-structured prob-
lems.” The nature of expertise, M. T. H. Chi, R. Glaser, and M. J. Farr,
eds., Lawrence Erlbaum, Hillsdale, NJ.
Wijekumar, K. J. (2001). “The effect of and effect with computer tools
in experts’representation of ill-structured problems.” Diss. Abstr.
Int. A, 61(12-A), 2378.
Wilson, P. S., ed. (1993). Research ideas for the classroom: High school
mathematics, MacMillan, New York.
Zeng, Y. (2002). “Axiomatic theory of design modeling.” J. Integr. Des.
Process Sci., 6(3), 1–28.
Zeng, Y. (2004). “Environment-based formulation of design problem.”
J. Integr. Des. Process Sci., 8(4), 45–63.