Concept mapping techniques

in computing education
Mark Dorling
@MarkDorling / m.d.dorling@qmul.ac.uk
PhD Student
Cognitive Science Department

Concept maps in computing education

Example: Image representation

Dorling (2018)
Introduction
• Concept maps are one of a “broader family” of graphical techniques
(Hay, Wells and Kinchin, 2008)
• Broader family include: mind mapping (Buzan and Buzan, 2000) and
spider diagrams (Mintzes, Wandersee and Novak, 1997)
• Very specific rules that differentiate concept mapping from the other
techniques and tools (Novak, 1998)

Teaching & Learning Interventions in Computing

Background
• Joesph Novak as being the creator of concept maps in 1970’s (Novak
and Gowin, 1984)
• Developed as part of his research program at Cornell University
(Novak and Musonda, 1991)
• Developed to find better ways for learners’ to illustrate and represent
their conceptual understanding, including the structure of that
understanding (Mintzes, Wandersee, & Novak, 2001) in the Science at
an early age (Novak and Musonda, 1991; Novak, 1990)
What are they?
• Concept maps are for organising, structuring, and representing
knowledge (Novak and Canas, 2006)
• Concept maps provide a tangible way to display how your
mind "sees" a particular topic. (Buzan and Buzan 2000; Cañas, Novak
and Reiska, 2012)
• Concept maps are a two dimensional representation (Mintzes,
Wandersee and Novak, 2001)
Learning theory
• Novak’s research program was based on David Ausubel’s learning
psychology and theory of Assimilation (Novak and Musonda, 1991)
• Ausubel’s theories have commonality with those that involve schema
as a central principle, for example, Bartlett, and influenced the work
of Piaget and constructivism i.e. assimilation and accommodation
(Beyerbach and Smith, 1990)
• “Meaningful learning” occurs when new information is linked to
existing knowledge (relevant concepts anchors) in the learner’s
cognitive structures (Ausubel 1963)

Concept maps in computing education

Meaningful learning…?
• Concept maps most effective when used as part of the learning
process (Anderson-Inman and Zeitz, 1993)
• ‘Proximity’ between learners’ prior learning and target knowledge
i.e. planning (Novak, 1998; Hay, Wells and Kinchin, 2008)
• Understanding learners prior knowledge i.e. misconceptions and gaps
in knowledge (Mason, 1992; Mintzes, Wandersee and Novak, 2005;
Novak, 1990; Novak and Gowin, 1984; Roberts, 1999; Kinchin, 2000)
Why?
• Conceptual proximity (difficulty in interconnecting and interrelating
concepts that are widely separated in classes or a passage of text)
(Fisher et al., 1990)
• Learners’ converting knowledge structures delivered in linear lessons
into a hierarchical (or similar) structure and back again when being
examined using traditional methods of assessment (Kinchin, 2016)
• There is a need to support educators with learning more meaningfully
so that they can transfer these learning styles to their learners (Hoz,
Tomer and Tamir, 1990; Beyerbach and Smith, 1990)
Why concept mapping…?
• Concept maps are a graphical organiser that aids leaners in bridging
the new and existing concepts (Novak and Canas, 2006)
• When learners create concept maps, they are forced to integrate new
information into their existing knowledge structures (Keppen and
Hays, 2008),
• Learners’ are encouraged to engage in the process of meaningful
learning by construct meaning for themselves through concept
mapping techniques (Keppens and Hay, 2008)
• Various studies have shown increased motivation in learners
compared to traditional forms of assessment (Edward and Fraser
(1983; Fisher et al., 1990; Horton et al., 1993; Markow and Lonning,
1998; Johnstone and Otis, 2006)
Our motivations
• Constructing a concept map is a demonstration of Bloom et al. (1956)
taxonomy of higher order cognitive performance: [Analyse], Evaluate
and Synthesis (Edmondson, 2000)
• Novak (1990) suggests that using concept mapping techniques with
pre-service and in-service educators could have two potential
benefits:
1. Helping educators with changing their own learning styles away
from rote learning toward more meaningful approaches, seeking to
make their own conceptual understanding more transparent;
2. They will become skills and confident in the use of concept maps
and may use these techniques with their learners’ in the classroom.
Why now…?
• Can support learners’ of all ages (from primary education through to
undergraduate level studies and Initial Teacher Training)
• In a range of subjects including Mathematics and the Sciences
• Where educators apply constructivist pedagogical approaches.
• Historical pedagogical trend in teaching Computing concepts, principles
and computational thinking has strong roots in the constructivist and
constructionist pedagogical approaches (Sentance and Csizmadia 2015,
2016)
• Making pedagogical judgements and being able to differentiate a resource
to meet the learners’ needs are dependent upon the teacher’s own subject
knowledge (Sentance and Csizmadia, 2016).
Approaches

Concept maps in computing education

Activities
• Ruiz-Primo and Shavelson (1996) and McClure, Sonak and Suen
(1999) suggest that concept mapping is compose of two component
activities:
1. Learner produces concept map:
a. A task is set;
b. A response is generated;
2. Educator evaluates the concept maps produced using a scoring
method.
Approaches
Chang, Sung, and Chen (2001) labelled concept mapping
approaches as either:

Construct-on-scaffold Construct-by-self
• Learners using pre-constructed • Learner constructed and
concept maps modified concept maps

Concept maps in computing education

Construct-on-scaffold
• Construct-on-scaffold styled concept mapping activities reduce
cognitive load for learners because the structuring of concepts and
relationships has been done by someone else (Paas, 1992; Sweller,
1998)
• Learners’ are presented with a:
• List of concepts and/or words for the linkages to use (Markham,
Mintzes and Jones, 1994; Osmundson et al., 1999)
• Predefined concept map structure created by an expert (educator)
and they are asked to fill in the blanks (Zeilik et al.,1997; Coleman,
1998)
Construct-by-self
• The value is in learners developing their own cognitive structures of
the given domain and representing them in a concept map (Kinchin,
2016)
• Construct-by-self style mapping approaches offer the greatest
learning potential (Canas, Novak and Reiska, 2010)
• Cognitive value to learners is in the process of developing their own
conceptual models and having the opportunity to revise their models
as they assimilate new knowledge (Stanton, Taylor and Tweedie,
1992; Zeiliger, Reggers and Peeters, 1996)
• Both approaches are applicable depending upon what the educator
hopes to achieve from using the approach (Canas, Novak and Reiska,
2012).
Comparison of approaches
• Construct-by-self style mapping approaches offer the greatest
learning potential (Canas, Novak and Reiska, 2010)
• Cognitive value to learners is in the process of developing their own
conceptual models and having the opportunity to revise their models
as they assimilate new knowledge (Stanton, Taylor and Tweedie,
1992; Zeiliger, Reggers and Peeters, 1996)
• Both approaches are applicable depending upon what the educator
hopes to achieve from using the approach (Canas, Novak and Reiska,
2012).
Construct-by-self

Concept maps in computing education

Structure
• Most concept maps represent a hierarchical structure, with the
overall, broad concept first with connected sub-topics, more specific
concepts, following.
• Concept maps begin with a main idea (or concept) and then branch
out to show how that main idea can be broken down into specific
ideas.
• Very specific rules that differentiate concept mapping from the other
techniques and tools (Novak, 1998)
• A five-step process for successfully constructing a concept map
(Novak, 1998; Novak and Canas, 2006; and Novak 2007)
Teaching & Learning Interventions in Computing
Concepts
• A concept can be perceived as an idea, principle, object or event
(process) (Fisher, 1990).
• The concepts given a label, usually specialist vocabulary (Fisher et al.,
1990)
• In a concept map, concepts are usually placed in an enclosed circle,
oval or square shape (Novak, 1998)
• Although, concept labels are usually represented using a word, phase
or symbol but can be illustrations (Stoica, Moraru and Miron, 2011)
• Specific examples of the concept can be added to the concept map
but these are not usually encased in a shape (Novak and Canas, 2006)

Teaching & Learning Interventions in Computing

Relationships
• Relationships between two concepts can be represented by
connecting a line between the concepts is called a linkage (Novak,
1998)
• Concepts are connected to other concepts boxes using a line with an
arrow (Novak and Canas, 2008)
• The linkages between concepts often have words or phases
connecting them (Novak and Canas, 2007)
• These linkage descriptions rarely use the technical vocabulary and
usually use “ordinary language” typically expressed as verbs (Fisher et
al., 1990)
Teaching & Learning Interventions in Computing
Propositions
• When two or more concepts connect (concept–relationship–concept)
then this is called a “proposition” (Nesbit and Adesope, 2006) or an
“instance” (Fisher, 1990)
• Important concepts will have lines to and from several other concepts
generating a network.
Cross-link
• A ”cross-link” is when a connection
between:
• concepts in different areas of the same
concept map is made;
• concepts in different concept maps are
made (Stoica, Moraru and Miron (2011)
• In identifying cross-links the learner is
making a “creative leap” (Novak and
Canas, 2008)
Buhmann and Kingsbury (2015)
5-step process to creating one…

Concept maps in computing education

Step 1: Identify
1. Identify concepts that you think are in anyway associated with the topic.
Do this from:
a) Memory;
b) Your notes;
c) Instructional materials;
2. Suggestion: In text, circle concepts and underline relationships;
3. Make a list of all the concepts. Describe each concept as briefly as
possible; usually one or two words per concept will be suffice. Keeping
your descriptions concise will prevent your map from becoming bloated
and text-heavy;
4. Depending upon the topic, a concept map usually has 15 – 25 concepts.
Evaluate the hierarchal list of concepts, reduce the number by focusing
on the most important;

Concept maps in computing education

Step 2: Organise (structure)
1. Transfer the list of concepts on to small Post-It notes, one concept per
note, then stick the post-it notes (concepts) on to a large sheet of plain
white paper. Suggestion:
a) Use one post-it note colour for notes and/or instruction and a different colour for
memory
2. Place each of your concepts into a hierarchical order as an actual list or
arrange (lay) them in the approximate order in which you think they
appear in your concept map;
a) The most general Concepts should be at the top
b) The most specific concepts and examples at the bottom of the list or concept map:
c) Within your list or concept map, try to create groups and sub-groups of related
concepts;
3. Introduce new items that you initially omitted in step 1;

Concept maps in computing education

Step 3: Represent (Layout)
1. Try and rearrange the concepts (post-it notes) into an arrangement
that reflects your understanding of the topic;
2. Use a consistent hierarchy in which the most important concepts at
the top;
3. Within sub-group, place closely related items near to each other
ready for step 4;
4. Expand your concept map one level at a time.

Concept maps in computing education

Step 4: Links (connections)
1. Draw lines with arrows to show the relationship between
connected concepts;
2. Try to label these connecting lines to define the relationship using a
word or short phrase e.g. specific verb like “provides” or
“encompasses”
3. Many arrows can originate or terminate on particularly important
concepts.
4. If possible, add cross-links to other domains.

Concept maps in computing education

Step 5: Revise
1. Feedback:
a) Getting feedback: Find someone to listen to you whilst you describe the
map out loud;
b) Giving feedback: Focus feedback on how well the explanation is reflected in
the actual concept map.
2. Ask these questions of your concept map:
a) Are there any important concepts or relationships missing?
b) Are the concepts and relationships correct?
c) Can I use a more accurate linking word to represent this relationship?
d) Is there a better position for this concept or group of concepts?
3. Make revisions to your concept map;

Concept maps in computing education

Practice scenario

Concept maps in computing education

Concept maps in computing education
What makes a good concept map?
• There is no agreement in the research literature on what makes a good concept map
(Derbentsevaet, Safayeni and Cañas, 2004)
• Novak and Canas (2008) stated seven structural characteristics for constructing a good
concept map:
1. Clarity in the focus question;
2. Use of boxes or circles for each concept;
3. Hierarchy of main concepts;
4. Whilst constructing the concept map identify the key concepts;
5. Relationships between concepts are shown (to form appropriate propositions);
6. Directional arrows are added on all relationship links between concepts (propositions);
7. Connecting different domains using cross-links.
• Chin (2001) also highlights the importance of learners providing (if applicable) specific
examples at the bottom of branches in concept maps.
Example of a good concept map

McClure, Sonak and Suen (1999)

 Concept map structures

Kinchin, Hay, and Adams (2000)

Buhmann and Kingsbury (2015)

What makes an excellent concept map?
• Cañas Novak and Reiska (2015) describes five characteristics that
distinguishes a good concept map from an excellent concept map:
1. Concise with clear focus on scope;
2. Clarity and elegance in communicating an idea;
3. Explanatory of an idea;
4. Balanced in structure;
5. Appropriate for audience.
• The structure of the concept map should match the content for the
domain being represented (Derbentsevaet, Safayeni and Cañas, 2004;
McDaniel, Roth and Miller, 2005).
Example of a Excellent concept map

Buhmann and Kingsbury (2015)

Computing scenario

Concept maps in computing education

Avoiding pitfalls
• Quality of the question or problem set can affect the eventual quality
of the concept map constructed by learners (Derbentsevaet, Safayeni
and Cañas, 2004; Novak and Canas, 2006)
• Learners’ go off task by producing a general domain concept map not
addressing the question of problem set (Novak and Canas, 2006)
• Trelease (2014) framed the question as “What are the key concepts
in…?”
Avoiding pitfalls
• Limit their number of concepts to about 20 and no more than 30
(White and Gunstone, 1992; Novak 1998; Anderson and Huang, 1989;
Rice, Ryan and Sampson, 1998; Novak and Canas, 2006; Novak, 2007)
• Limiting the number focuses learners on the most important concepts
as the learner perceives it and when working in diagrammatic form
reduces the load on working memory (Stensvold and Wilson, 1992)
Formative vs. summative
• Learners’ constructing a concept map it can only been seen as a
“snap-shot” (“single observation”) of their understanding at that
moment in time (McClure, Sonak and Suen, 1999; Kinchin, 2012)
• Learners’ knowledge is continually developing therefore concept
maps should always be seen as work in progress not the final product
(Wexler, 2001; Canas and Novak, 2008; Kinchin, 2016)
• Use as a formative assessment tool when making judgements about
learner progress (McClure, Sonak and Suen, 1999)
• Kinchin (2001) offers a note of caution, that quantitative assessment
methods that calculate a quantitative scores over a period of time to
measure improvement is ultimately “unrevealing”.