Novak 2021 Mathematical Modeling
Novak 2021 Mathematical Modeling
Novak 2021 Mathematical Modeling
net/publication/356642039
CITATIONS READS
2 587
1 author:
Elena Novak
Kent State University
45 PUBLICATIONS 799 CITATIONS
SEE PROFILE
All content following this page was uploaded by Elena Novak on 30 November 2021.
Elena Novak
Abstract
Mathematical modeling describes how events, concepts, and systems of interest behave in the
world using mathematical concepts. This research approach can be applied to theory construction
and testing by using empirical data to evaluate whether the specific theory can explain the
empirical data or whether the theory fits the data available. Although extensively used in the
physical sciences and engineering, as well as some social and behavioral sciences to examine
theoretical claims and form predictions of future events and behaviors, theory-oriented
mathematical modeling is less common in educational technology research. This article explores
the potential of using theory-oriented mathematical modeling for theory construction and testing
in the field of educational technology. It presents examples of how this approach was used in
social, behavioral, and educational disciplines, and provides rationale for why educational
1
1. Introduction
Research in the field of educational technology and design has often been criticized for
having little impact on improving learning and contributing to basic theories that inform teaching
(Cilesiz & Spector, 2014; Reeves, 2006; Seel, 2009; Suppes, 1978), being methodologically
weak (Bulfin, Henderson, Johnson, & Selwyn, 2014) and under-theorized (Jones & Czerniewicz,
2011; Markauskaite & Reimann, 2014). For example, a recent analysis of educational technology
research of over 500 empirical articles found that few papers engaged or informed theory; most
papers either did not mention theory at all or provided vague references to theory (Hew, Lan,
Tang, Jia, & Lo, 2019). The authors argued about the importance of using existing theories to
frame empirical research, because they allow for findings to be generalized across different
It is important to note that the term ‘theory’ means different things to different people and
there is no consensus about what theory is. For example, Simon and Newell (1956) proposed to
“define a theory simply as a set of statements or sentences” (p. 67). Other scholars viewed theory
as a set of statements and propositions that explain the relationship(s) of concepts or variables
(Hollander, 1967; Jaccard & Jacoby, 2009). Merton (1968) distinguished between middle-range
theories and unified theories: “[middle-range] theories that lie between the minor but necessary
working hypotheses that evolve in abundance during day-to-day research and the all-inclusive
systematic efforts to develop a unified theory that will explain all the observed uniformities of
social, behavior, social organization and social change" (p. 39). Opp (1970) refined Merton’s
definition of a middle-range theory as a theory that can "explain a certain class of particular
('sociological') events" (p. 245) and of a unified theory as "a theory from which at least all
existing theories of the middle range can be deduced in their original or in modified form" (p.
2
245). He argued that “the unified theory is empirically testable and has an extremely high
explanatory power. Otherwise it could not explain the particular events and theories which it is
Constructing new theories and testing existing ones is considered among the most important
scientific efforts that advance science (Kuhn, 1996; Mintzberg, 2005; Morgan & Wildemuth,
2009). A growing number of researchers advocate the use of mathematical modeling (also
theoretical foundations of the field and improving empirical research (Jaccard & Jacoby, 2009;
Rodgers, 2010). Mathematical modeling describes how events, concepts, and systems of interest
behave in the world using mathematical concepts. It is a powerful research tool that has been
and form predictions and forecasts of future events and behaviors. However, development and
utilization of mathematical models is more common in the physical sciences and engineering
than in the social and behavioral sciences (Jaccard & Jacoby, 2009). Within the realm of the
social and behavioral sciences, some disciplines like economics, for example, are also known for
their extensive use of mathematical modeling as a research and policy tool. In education and
educational technology fields, however, mathematical modeling is less common, and its potential
is yet to be recognized.
According to Pearl (2000, p. 202), model is “an idealized representation of reality that highlights
3
some aspects and ignores others.” A mathematical model can be described using a mathematical
variables. The goal of mathematical model building is to develop a model that will provide an
accurate, yet relatively simple representation of reality (Figure 1). Social and behavioral
scientists usually do so by using regression modeling that can predict the impact of the
intervention using desired moderator variables. Of course, many would rightly argue that human
behaviors are complex and cannot be captured using simple linear relationships. In mathematical
model building, this tension is often addressed with the use of appropriate statistical methods.
Theory
Real-world problem Conceptual model
Identify/construct
mathematical model (e.g.,
equations, graphs)
Fig1. An elementary view of mathematical modeling that shows the phases in building a
mathematical model and how models relate to real-world problems (inspired by Dym, 2004, and
Jaccard & Jacoby, 2009)
Although variables are critical to many theories, not all theoretical approaches use variable-
oriented approaches to develop theories. These theories use a process-oriented approach and/or
4
narratives to explain a specific phenomenon (Jaccard & Jacoby, 2009) and mathematical
modeling may not be an appropriate approach for these theoretical frameworks. For instance,
instead of defining gender as a variable with two levels, female and male, these theories focus on
approaches that rely on variables to explain different relationships among specific concepts, such
as explanatory (“Y because of X”) or design theories (“to achieve Y, do X”) can use
mathematical modeling to construct and test the theories (Malone, 1985). Such approaches rely
on the concept of causality to identify systematic relationships between variables that can explain
our environment, organize constructs of interest, and predict future events. Typical model
construction involves the following steps (Boland, 2014; Jaccard & Jacoby, 2009):
1. Identify exogeneous and endogenous variables: The process begins with identifying
variables that explain the causal relationship(s) of interest as well as basic behavioral
relationships between the variables. In particular, the modeler needs to identify which of those
variables are exogeneous and which are endogenous. Variable X is exogeneous if a change in X
causes changes in Y. That makes variable Y endogenous, or the variable that can be explained by
the model. After identifying the variables, the modeler needs to consider how to measure the
2. Include model assumptions: Boland (2014) argues that during the second phase of model
well as assumptions about the relationships between the identified variables and their parametric
specifications.
5
3. Identify possible functions for model construction: The modeler needs to identify possible
functions that can capture the relationship between the variables of interest. Toward this end, a
wide variety of functions can be employed including, but not limited to linear, exponential,
logarithmic, and polynomial functions. Although linear relationships are common in modeling,
they do not always explain all possible relationships between quantitative variables. For instance,
the relationship between levels of test anxiety and test performance takes a form of a nonlinear
inverted-U function (a U drawn upside down) where low levels of anxiety are beneficial for test
performance but high levels of anxiety are detrimental to a person’s success on the test, as
tension and worries consume critical cognitive resources that could otherwise be directed to the
4. Collect data to examine the model performance: Once the mathematical model is
constructed (e.g., equations, graphs), it is time to collect empirical data to examine the model
performance. Specifically, the modeler needs to explore how well the values of the endogenous
variable(s) fit the real-world values. If the degree of fit is unsatisfactory, the original model can
5. Revise the model and evaluate its performance with new data. After making the necessary
adjustments to the original model, a new set of empirical data should be collected to evaluate the
revised model’s performance. If the revised model is conceptually plausible and performs well,
process of developing and testing mathematical models. To discuss it in detail would be too
extensive for the purpose of this paper. Interested readers may find Jaccard and Jacoby’s (2009)
6
book “Theory Construction and Model-Building Skills” helpful for learning about the cognitive
heuristics, tools, and approaches for model construction. Moreover, structural equation modeling
can be used for mathematical modeling as well (Jaccard & Jacoby, 2009).
The basic idea behind mathematical model building is to “translate the verbal problem into a
mathematical concepts. But how are theories different from models? Theories explain certain
phenomena and critical explanatory variables or mechanism, but generally are “flexible about the
expected conclusions of empirical research” (Nelson & Winter, 1974, p. 886), while models
Roald Hoffman (2003), a Nobel Prize-winning scientist, in his commentary about theories
and models compared the process of a theory development to storytelling: “When things are
complex yet understandable, human beings weave stories” (p. 224), and “Many theories are
popular because they tell a rollicking good story, one that is sage in capturing the way the world
works, and could be stored away to deal with the next trouble” (p. 225). His view on the
importance of models is that “relatively uncomplicated models … play a special role in the
acceptance and popularity of theories among other theorists… The models become modules in a
theoretical Erector set, shuttled into any problem as a first (not last) recourse. … taking one piece
of experience over to another” (p. 225). That is to say that mathematical modeling is clearly
7
Earlier attempts in the late 1930s to represent behavioral theories using mathematical
equations were done primarily for theoretical purposes only (see Boland (2014) for an extensive
overview of model building in economics). Today, models are designed with a specific goal in
mind, usually as a research or policy instrument. Such models are called empirical models.
designed to identify patterns in the available data in order to form short-term forecasts or
evaluate the impact of possible policy changes, without drawing its inferences from a specific
theory. Theory is used, but only as it fits the data of interest (Boland, 2014). Theory-oriented
empirical models are designed to test a theory or theoretical claims of interest. This type of
model building utilizes empirical data to evaluate whether the specific theory can explain the
cannot support a theory in question does not imply that the theory is wrong, -- also known as the
Duhem-Quine thesis in the philosophy of science. It only suggests that the theoretical statement
is not confirmed (not to be confused with ‘disconfirmed’) or that the model builder made false
background assumptions in the model design, e.g., assuming a linear relationship between
variables instead of a quadratic one, or using tools/instruments that produce measurement errors.
Morgan and Wildemuth (2009) wrote: “A theoretical model is not defeated by the discovery of
exceptions to the theory, but only by the evolution of a better model, or at least one that better
answers the questions that scientists are asking at a given time” (p. 42). Kuhn (1996) notes that
theory testing and the discovery of anomalies is central to developing a scientific understanding
about our world, as “theory works only until a critical mass of anomalies and exceptions to it are
8
found, at which point, it becomes untenable and must be replaced by a new theoretical model”
There are many challenges to implementing mathematical modeling in education, social, and
behavioral disciplines due to the complex, dynamic, unique, and mostly obscured nature of
reality (Jaccard & Jacoby, 2009), lack of appropriate methodological data analysis approaches,
and high levels of uncertainty (Dabbaghian & Mago, 2014). Nevertheless, Rodgers (2010)
argues that mathematical modeling should be used more extensively in behavioral research and
graduate training to build and evaluate various behavioral models: “Researchers should be
developing mathematical models and evaluating them statistically” (p. 5). Yet, it has not been
Mathematical models are prominent in social and behavioral sciences for analyzing human
decision making. They are also used in theories of memory, learning, language and in the
analysis of social networks, such as institutions, communities, and trade networks. For instance,
mathematical modeling was used to study short-term and long-term memory (Atkinson &
Shiffrin, 1968), musical perception (Shepard, 1982), human reproduction describing childbearing
decision making (Rodgers & Doughty, 2001), onset of various social activities, like smoking and
drinking (Rodgers, 2003; Rowe & Rodgers, 1991) and sexual behavior (Rodgers & Rowe, 1993).
In political science, mathematical models are prominent in analyzing political behavior such as
Among other benefits of using mathematical modeling in research is its ability to address the
causal generalization problem in research that evaluates the causal impacts of educational
interventions in order to inform policy decisions, such as whether to integrate a new educational
9
technology into a curriculum (Rodgers, 2010; Scheerens & Blömeke, 2016; Shadish, Cook, &
Campbell, 2002; Tipton & Olsen, 2018). Oftentimes, the results from the study do not generalize
to the population that would be impacted by these decisions, because the impact of an
intervention varies across educational settings and learners. This is one of the reasons why
educational technology research had very little impact on improving instructional practices and
student learning, because we cannot adapt our instructional interventions to the myriad variables
that surround and interact with any given instance of instruction (Cronbach, 1975; Reeves, 2006;
Seel, 2009). Using a model-testing approach can address the causal generalization problem
Nevertheless, it is important to note that mathematical modeling is not the only approach for
improving the quality and impact of educational research. Other types of inquiry such as design-
based or qualitative research are used extensively to generate new knowledge, construct theories,
and devise solutions to complex educational problems. For instance, one of the important goals
of design-based research is to generate new knowledge that can inform theory development
(McKenney & Reeves, 2014; Puntambekar, 2018). Those developed theories can then be further
Many books have been written about theory construction and modeling as these ideas are at
the core of the scientific process. However, few of those introduce the topic to students or
professionals in education, social and behavioral sciences, who are new to theory construction
and modeling. This shortage of literature on teaching theory construction and modeling reflects
the curricular trends of many graduate programs in these disciplines. Graduate students in
education, social and behavioral sciences usually take several courses in research methods and
data analysis, but seldom receive extensive training in model building and theory construction
10
(Jaccard & Jacoby, 2009; Rodgers, 2010). Unsurprisingly, we do not see many applications of
This article presents examples of model-testing from social, behavioral, and education
The first example demonstrates the construction and testing of a unified theory in educational
technology. Keller (2008) developed the integrative theory of Motivation, Volition, and
reference to take a broad view of the learning, motivational, volitional, psychological, and
environmental factors that affect human performance. This theory can be characterized as a
macro theory that incorporates various micro theories, including information-processing theories
(Atkinson & Schiffrin, 1971; Mayer, 2001), motivation (Keller, 1999), action control theory
1999). Keller’s motivational theory (1999), also known as the ARCS model, served as one of the
foundation blocks that led to the development of the integrative theory of MVP. The ARCS
model describes motivational factors that influence learning and performance. Specifically, the
first three factors, i.e., attention, relevance, and confidence, relate to learning conditions. The
fourth ARCS factor, satisfaction, describes the product of the learner’s mental appraisal of their
11
ability to apply the newly acquired skills and knowledge to everyday life, which can be viewed
as a learning outcome.
systematically represent relationships among explanatory components of the MVP theory (Figure
2). The conceptual MVP model describes inputs (i.e., considered to play an important role in
explaining the theoretical principle), interactions among them, and outputs (i.e., measurable
behaviors to be explained). This conceptual model was later statistically tested using four
different data sets from distinct digital learning environments (electronic textbooks, simulations,
video games), learning content (statistics, biology, economic theory), settings (classroom, lab),
and learners (undergraduate and graduate students from various majors; Huang et al., 2010;
Fig2. The MVP model of motivation, volition, and performance according to Keller (2008a).
Reprinted with the permission of John Keller from “An Integrative Theory of Motivation,
Volition, and Performance,” by J. Keller, 2008, Technology, Instruction, Cognition and
Learning, 16, 79-104, p. 94.
12
The first investigation in this area revealed a significant model that provided empirical
support for the ARCS model in an online instructional game with undergraduate students from
various majors (Huang et al., 2010). This mathematical model described the relationship between
the motivational (attention, relevance, and confidence) and outcome processing (satisfaction)
Later, Novak (2014) used Huang et al.’s (2010) model (Equation 1) to develop an initial
mathematical MVP model and statistically evaluate it using motivational, volitional, and
performance data with graduate students in a digital simulation-based learning environment. The
+ D × Volition(s) (2)
Recently, Novak, Daday, and McDaniel (2018) statistically evaluated the initial MVP model
(Equation 2) for (a) fitting two new data sets with undergraduate biology from two different
technology-rich learning environments, and (b) expanding the model by including additional
variables that affect student outcome processing in e-text learning environments. Their study
confirmed the generalizability of the initial MVP model across different digital learning
environments and identified four additional variables in the MVP model, i.e., intrinsic cognitive
load (IL), extraneous cognitive load (EL), number of academic hours earned in the degree
13
Satisfaction (s) = A × Attention(s) + B × Relevance(s) + C × Confidence(s) + D × IL(s) +
(A, B, C, and D are positive parameters; E, F, and G are negative parameters; s = student)
A comparison of MVP model’s variables across the studies that developed and tested
MVP models showed that the model produced consistent empirical support for the elements of
the MVP theory across various technology-rich learning environments and student populations
(Huang et al., 2010; Novak, 2014; Novak et al., 2018). The MVP model has research and
practical implications. For researchers, it can explain how well the MVP theory matches the
reality of the learning, motivational, volitional, psychological, and environmental factors that
influence learning and performance. For practitioners, it can serve as a tool for diagnosing and
The second example illustrates the development and testing of a unified theory of user
acceptance of new technology. User acceptance of new technology is one of the most mature
research areas in the information systems field, and the literature offers myriad competing user
acceptance models and constructs. Oftentimes, researchers chose a “favored model” or had to
“pick and choose” related constructs (Venkatesh, Morris, Davis, & Davis, 2003), which did not
contribute to expanding theoretical foundations of the field. To address this challenge, Venkatesh
et al. (2003) developed the Unified Theory of Acceptance and Use of Technology (UTAUT) to
create a unified view of the information technology acceptance research that explores user
14
Venkatesh and colleagues empirically compared various user acceptance models to identify
conceptual and empirical similarities across the models that were later used to create and
empirically validate a unified model. The authors empirically tested the UTAUT in four
different professional settings and cross-validated it with data from two additional organizations.
This empirical testing revealed variables that influenced usage intention and usage behavior as
well as several additional variables that mediated these relationships (Figure 3).
Performance
Expectancy
Effort
Expectancy
Behavioral
Use Behavior
Intention
Social Influence
Facilitating
Conditions
Fig3. UTAUT Research Model. Adapted from “User acceptance of information technology:
Toward a unified view,” by Venkatesh et al., 2003, MIS Quarterly, 27(3), p. 447.
To examine the impact of the UTAUT on the field, Venkatesh, Thong, and Xu (2016)
conducted a comprehensive review of the UTAUT literature eleven years after the initial
development of the UTAUT. Their review revealed that the UTAUT was used extensively across
15
a variety of settings and users, including schools, hospitals, manufacturing and service sectors,
digital learning contexts, e-government services, repeatedly validating the relationships found in
the original UTAUT model. A total of 37 different UTAUT extensions were developed based on
the original UTAUT model (Figure 4). These UTAUT extensions extended the baseline UTAUT
model with new mechanisms and variables predicting technology use and/or behavioral
Performance
Expectancy
Effort New
Expectancy Behavioral Technology Outcome
Intention Use Mechanism
s
New
Social Influence
Exogenous
Mechanisms
Notes:
1. Boxes and arrows with
Facilitating
dashed lines represent
Conditions UTAUT extensions.
The effect of teacher education programs on future teachers’ knowledge and teaching
practices is one of the major research areas in the field of teacher education (Qian & Youngs,
2016). Research on teacher education clearly shows that teacher quality is the most important
school-related factor that affects student achievement (Goe, 2007; Qian & Youngs, 2016).
16
Blömeke and Kaiser (2011) used a model-testing approach to examine curricula of teacher
education programs in mathematics across 15 countries. Their research was motivated by a Tatto
et al. (2008) study that compared teacher’s mathematics content and pedagogical knowledge at
the end of their training from 15 countries in Africa, America, Asia, and Europe. Tatto et al.
(2008) found considerable differences of about one standard deviation in teachers’ knowledge
outcomes across the countries. These findings prompted Blömeke and Kaiser (2011) to examine
whether prospective teachers in different countries had similar opportunities to learn (OTL)
mathematics, mathematics pedagogy, and general pedagogy during their training. They grounded
(McDonnell, 1995), noting that research on OTL in teacher education is weak and not supported
by robust empirical data. The model-testing approach allowed Blömeke and Kaiser (2011) to
identify similar and distinct features of the curriculum of teacher education programs (OTL
profiles) across the countries and examine the relationships of these OTL profiles to cultural
contexts. In addition to laying a foundation for the discussion about homogeneity in mathematics
pedagogy, this approach helped pinpoint culturally shaped visions of teacher knowledge and the
teaching profession across different countries and within the countries’ programs.
Scheerens and Blömeke (2016) examined the relationships among teacher candidates’
learning in teacher education programs and PreK-12 student achievement using a modeling
approach. They created and tested an integrated theoretical model of school effectiveness and
teacher education effectiveness to better understand the processes mediating the transformation
of teacher knowledge into instructional quality that influences student achievement. Scheerens
17
and Blömeke (2016) argued that this “extended educational effectiveness model … enriches a
systemic interpretation of key levers of educational effectiveness and opens up black boxes at the
system and the classroom level” (p. 70) by allowing the systematical examination of the
relationships among different variables affecting teacher education and school effectiveness.
using technology – has been studied primarily in the field of human-computer interface.
Bessière, Newhagen, Robinson, and Shneiderman (2006) developed a theoretical model for
hamper goal achievement. They explored how computer frustration affects users’ experiences
associated with negative computing experiences focusing on two types of outcomes: (1)
situational factors that relate to specific events, e.g., the importance of the task, severity of the
interruption, and frequency of occurrence, and (2) dispositional factors, e.g., user self-efficacy,
as students’ experiences with technology directly relate to their attitudes toward the technology
and intentions to continue using the technology as a learning tool. Novak, McDaniel, Daday, and
Soyturk (2021) examined the application of Bessière and colleagues’ (2006) model of computer
frustration to study students’ frustration with e-textbooks. Novak et al.’s (2021) study pinpointed
18
presented in the Bessière and colleagues’ (2006) original model but also expanding it by
identifying new cognitive and motivational variables that influence e-text frustration.
TASK
Goal Interruption
Level of Goal
Computer
DISPOSITIONAL MEDIATORS
Commitment/
SITUATIONAL MEDIATORS
Importance Experiences
Severity of Psychological
Interruption/ Time Factors/ Self Efficacy
Loss
Mood
Strength of Desire/
Anticipation
Expectations
Fig5. The computer frustration model. Adapted from “A model for computer frustration: The
role of instrumental and dispositional factors on incident, session, and post-session frustration
and mood,” Bessière et al., 2006, Computers in Human Behavior, 22(6), p. 949.
The field of online (also referred to as distance) education has seen a tremendous growth
in scholarship over the past several decades. Michael Graham Moore in the Preface to his and
William Diesl’s (2018) Handbook of Distance Education notes that this growth has contributed
to the “splitting of the field into component specialties” (Moore, 2018; p. xii), such as e-learning,
19
Internet learning, distributed learning, networked learning, blended learning, virtual learning,
Web-based learning, distance learning, and tele-learning. Each of the component parts of the
journals, and unique conferences and interest groups. Although this “splitting” and the breadth of
the field is common to any maturing field, it creates a problem – a lack of unified foundational
knowledge and theory that is readily accessible to all specialties of online/distance education
(Moore, 2018). Research in this area utilizes a variety of recognized theoretical foundations such
as Community of Inquiry (Garrison, Anderson, & Archer, 2000), Connectivism (Siemens, 2004),
(Harasim, 2012), Transactional Distance (Moore, 1997), and Independent Study (Wedemeyer,
1981). In addition, various learning theories from behaviorist, constructivist, and cognitive
schools of learning can be applied to online learning (Anderson, 2008). To integrate the work of
many theorists, Terry Anderson developed a model of online learning as a “first step toward a
theory in which the two predominant forms of online learning – collaborative and independent
study – are considered” (Anderson, 2008, p. 7; Figure 6). Moreover, he emphasized the utility
aspect of online learning (i.e., effective online learning): "We need theories of online learning
that help us to invest our time and limited resources most effectively" (Anderson, 2008, p. 46).
Although the proposed theory was incomplete, the model identified many important variables
20
Fig6. Anderson’s (2011) model of online learning. Adapted from “The theory and practice of
online learning,” Anderson, T., 2011, Edmonton, AB: AU Press, p. 49.
online education (Figure 7). Picciano’s model was developed based on an analysis of some of the
cognitive theories and their derivatives, e.g., Wenger and Lave’s (1991) Communities of
Practice, Wenger’s (1998) Situated Learning, Atkinson and Shiffrin’s (1968) Information
21
Processing, neuroscience research (Willingham, 2008), and Knowles’ (1998) Adult Learning
education but does not explain the relationships among them. Picciano (2017) concludes his
develop an integrated theory of online education. Mathematical modeling can help answer this
question by empirically testing the proposed theory of online education. It can also expand the
theory by examining additional factors that are critical to online learning. Moreover, it can help
Fig7. Picciano’s (2017) model for online education. Adapted from “Theories and frameworks for
online education: Seeking an integrated model,” Picciano, A. G., 2017, Online Learning, 21(3),
166-190, p. 182.
22
Means, Bakia, and Murphy’s (2014) framework for research on online learning (Figure 8)
may prove to be a useful tool when applying mathematical modeling to testing and/or
the effectiveness of online and blended learning (Means at al., 2013), Means and colleagues
(2014) proposed a framework for research on online learning that offers four possible
dimensions that characterize online learning: (1) context, (2) instructional design, (3)
Fig8. Means and colleagues’ (2014) Four dimensions of online learning. Adapted from
“Learning online: What research tells us about whether, when and how,” Means et al., 2014,
New York: Routledge, p. 9.
They argued: “Our ability to accumulate knowledge about the kinds of online learning
experiences that produce desired effects for specific kinds of learners under a given set of
circumstances would be greatly enhanced if every research report used these dimensions and an
agreed set of more specific features within each dimension to generate comprehensive
23
descriptions of the interventions they studied” (Means et al., 2014, pp. 8-9). Clearly, this
recommendation suggests a unified or integrated approach to research on online learning and can
In summary, this section provided one example of how mathematical modeling can
advance the field of online education. There are many other areas of educational technology that
can benefit from using mathematical modeling. For example, research on barriers to technology
integration in K-12 education (Ertmer, 1999; Hew & Brush, 2007) is continuing experiencing an
overwhelming growth in scholarship. But most studies in this field take a narrow approach when
examining possible barriers to technology integration focusing on teachers’ technical skills and
attitudes toward technology (Hew & Brush, 2007). This approach is dangerous, because it
presents a very limited view, without taking into consideration the environmental, curricular,
psychological, and content knowledge factors that affect teachers’ ability and willingness to use
theory of technology integration in K-12 education can advance the theoretical foundations of
6. Conclusion
The field of educational technology, learning, and design offers a great number of
middle-range theories. In such fields, it is advisable to attempt the construction and statistical
testing of unified theories that synthesize multiple middle-range theories (Opp, 1970; Rodgers,
2010). Since a unified theory can explain a set of middle-range theories, then it can explain at
least the same events as all middle-range theories that informed the unified theory. Thus, unified
theories can facilitate and speed-up the process of inquiry as well as create opportunities for new
24
research directions that would not have otherwise happen. Moreover, a unified theory in the field
of educational technology can serve as a platform for research across different research
expand the theoretical foundations of the field as well as develop new theories that offer a new
perspective of looking at phenomena. Studies that use a theory to frame their research and find
refinements to the theory are especially valued due to their contribution to research and theory
Although mathematical modeling is not the only approach to theory construction, it offers
a sophisticated way of thinking about relationships between variables that can benefit theory
construction and testing efforts in the field of educational technology. For example,
mathematical modeling can be used to expand theories examining relationships among different
variables that do not have extensive empirical support. Well-developed theories can be used to
develop and test mathematical models in culturally diverse settings. In addition, models can be
tested with different educational technologies and learners in different contexts and settings.
Models that produce consistent results are considered more robust (Johns, 2006; Venkatesh et al.,
2016). These theories and models are particularly suited for future research that aims to examine
possible theoretical extensions in new contexts and settings. Models that do not provide
consistent support for the theories, or do not generalize well to new contexts and learners,
present research opportunities that can expand theoretical foundations of the field.
25
References
Anderson, T. (2008). The theory and practice of online learning (2nd Edition). Edmonton, AB:
AU Press.
Atkinson, R. C., & Shiffrin, R. M. (1968). Human memory: A proposed system and its control
processes. In Spence, K. W., & Spence, J. T. The psychology of learning and motivation
Atkinson, R. C., & Schiffrin, R. M. (1971). The control of short-term memory. Scientific
Atkinson, R. C., & Shiffrin, R. M. (1968). Human memory: A proposed system and its control
motivation: Advances in research and theory (Vol. 2), 89-195. New York, NY: Academic
Press.
Bessière, K., Newhagen, J. E., Robinson, J. P., & Shneiderman, B. (2006). A model for computer
frustration: the role of instrumental and dispositional factors on incident, session, and
doi:https://doi.org/10.1016/j.chb.2004.03.015
learn in primary teacher education and their relationship to cultural context and
Boland, L. A. (2014). Model building in economics: Its purposes and limitations. New York,
26
Bulfin, S., Henderson, M., Johnson, N. F., & Selwyn, N. (2014). Methodological capacity within
Cilesiz, S., & Spector, J. M. (2014). The Philosophy of Science and Educational Technology
Dabbaghian, V., & Mago, V. K. (2014). Theories and Simulations of Complex Social Systems.
Ertmer, P. A. (1999). Addressing first- and second-order barriers to change: Strategies for
61.
Garrison, D. R., Anderson, T., & Archer, W. (2000). Critical inquiry in a text-based
Goe, L. (2007). The link between teacher quality and student outcomes: A research synthesis.
Washington, DC: National Comprehensive Center for Teacher Quality. Retrieved from
http://ncctq.learningpt.org/ publications/LinkBetweenTQandStudentOutcomes.pdf
27
Harasim, L. (2012). Learning theory and online technologies. New York: Routledge/Taylor &
Francis.
Hew, K.F., & Brush, T. (2007). Integrating technology into K-12 teaching and learning: Current
Hew, K. F., Lan, M., Tang, Y., Jia, C., & Lo, C. K. (2019). Where is the “theory” within the field
Hoffman, B. (2010). “I think I can, but I'm afraid to try”: The role of self-efficacy beliefs and
Hoffman, R. (2003). Why buy that theory? In O. Sacks (Ed.), The best American science writing:
Hollander, E. P. (1967). Principles and methods of social psychology. Oxford, UK: Oxford
University Press.
Holmberg (1985). The feasibility of theory of teaching for distance education and a proposed
theory (ZIFF Paiere 60). Hagen, West Germany: Fern Universitat, Zentrales Institute fur
ED290013).
Huang, W.-H., Huang, W.-Y., & Tschopp, J. (2010). Sustaining iterative game playing processes
28
Jaccard, J., & Jacoby, J. (2009). Theory construction and model building skills: A practical guide
Jones, C., & Czerniewicz, L. (2011). Theory in learning technology. Research in Learning
Keller, J. M. (1999). Using the ARCS motivational process in computer-based instruction and
distance education. New Directions for Teaching and Learning, 78, 39-48.
Knowles, M.S., Holton, E.F. & Swanson, R.A. (1998). The adult learner (5th Edition). Houston:
Butterworth-Heinemann Publishers.
Kuhl, J. (1987). Action control: The maintenance of motivational states. In F. Halisch & J. Kuhl
Kuhn, T. S. (1996). The Structure of Scientific Revolutions. Chicago, IL: University of Chicago
Press.
Markauskaite, L., & Reimann, P. (2014). Editorial: e-Research for education: Applied,
45(3), 385–391.
29
Means, B., Bakia, M., & Murphy, R, (2014). Learning online: What research tells us about
Means, B., Toyama, Y., Murphy, R., & Baki, M. (2013). The Effectiveness of Online and
Record, 115.
Mintzberg, H. (2005). Developing theory about the development of theory. In M. Hitt & K.
Moore, M. G. & Diehl, W. C, (2018). Handbook of Distance Education. New York: Routledge
Moore, M. G. (1973). Towards a theory of independent learning and teaching. Journal of Higher
Morgan, C., & Wildemuth, B. M. (2009). Questions related to theory. Applications of social
research methods to questions in information and library science (pp. 40-50). Westport,
Nelson, R., & Winter, S. (1974). Neoclassical vs. Evolutionary Theories of Economic Growth:
30
Novak, E. (2014). Toward a mathematical model of motivation, volition, and performance.
Novak, E., Daday, J., & McDaniel, K. (2018). Using a mathematical model of Motivation,
018-9599-5
Novak, E., McDaniel, K., Daday, J., & Soyturk, I. (2021). Understanding student frustration
Opp, K.-D. (1970). Theories of the middle range as a strategy for the construction of a general
Pearl, J. (2000). Causality: Models, reasoning, and inference. Cambridge, England: Cambridge
University Press.
Picciano, A. G. (2017). Theories and frameworks for online education: Seeking an integrated
Qian, H., & Youngs, P. (2016). The effect of teacher education programs on future elementary
31
Reeves, T. C. (2006). Design research from the technology perspective. In J. V. Akker, K.
London: Routledge.
Rodgers, J. L. (2003) EMOSA sexuality models, memes, and the tipping point: Policy and
Rodgers, J. L., & Doughty, D. (2001). Does having boys or girls run in the family? . Chance, 14,
8-13.
Rodgers, J. L., & Rowe, D. C. (1993). Social contagion and adolescent sexual behavior: A
295X.100.3.479
Rowe, D. C., & Rodgers, J. L. (1991). Adolescent smoking and drinking—are they epidemics?
Scheerens, J., & Blömeke, S. (2016). Integrating teacher education effectiveness research into
doi:https://doi.org/10.1016/j.edurev.2016.03.002
Seel, N. M. (2009). Bonjour tristesse: Why don’t we research as we have been taught?
Shadish, W., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental
32
Shepard, R. N. (1982). Geometrical approximations to the structure of musical pitch.
Siemens, G. (2004). Connectivism: A learning theory for the digital age. International Journal of
http://www.itdl.org/Journal/Jan_05/article01.htm
Simon, H. A., & Newell, A. (1956). Models: Their uses and limitations. In D. White (Ed.), The
Straub, D. W. (2009). Editor’s comments: Why top journals accept your paper. MIS Quarterly,
33(3), iii–x.
Suppes, P. (1978). Impact of research on education: Some case studies. Washington, DC:
Tatto, M. T., Schwille, J., Senk, S., Ingvarson, L., Peck, R., & Rowley, G. (2008). Teacher
Lansing, MI: Teacher Education and Development International Study Center, College of
Tipton, E., & Olsen, R. B. (2018). A Review of Statistical Methods for Generalizing From
doi:10.3102/0013189x18781522
Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of
33
Venkatesh, V., Thong, J. Y. L., & Xu, X. (2016). Unified theory of acceptance and use of
technology: A synthesis and the road ahead. Journal of the Association for Information
Wedemeyer, C.A. (1981). Learning at the back door: Reflections on non-traditional learning in
Wenger, E. & Lave, J. (1991). Learning in doing: Social, cognitive and computational
Zawacki-Richter, O., Bäcker, E. & Vogt, S. (2009). Review of distance education research (2000
https://doi.org/10.19173/irrodl.v10i6.741
Mahwah.
34