International Journal of Cyber Society and Education

Pages 99-116, Vol. 3, No. 2, December 2010

The Development of Online Learning Styles Inventory: An

Exploratory Study
Yu-Ching Liu
Hsuan Chuang University
E-Mail: [email protected]
Yu-Chih Doris Shih
Fu-Jen Catholic University
E-Mail: [email protected]
Ryh-Wu Yeh
Hsuan Chuang University
E-Mail: [email protected]

ABSTRACT
This study intends to develop an online learning styles inventory. This paper reports
the preliminary result of the development. We first developed 168 items, 105 items were
selected for pilot study, and finally 64 items were chosen for the main study. The
procedure of developing this inventory included expert evaluation, factor analysis,
internal consistency reliability, and test-retest correlation. In order to improve the
reliability and validity of the subscales, continuous empirical study is required.
Keywords: Learning Style, Inventory Development, Online Learning, Distance Education
INTRODUCTION
With the rapid advance in telecommunication techniques, many countries have
adopted online learning (which is also referred to as e-learning) into school education and
life-long learning, offering more diverse and convenient learning options for students.
College institutions in Taiwan have followed such trend, and established various distance
learning courses. However, is it suitable for all college students to take online learning

International Journal of Cyber Society and Education 100

courses? According to some related studies, learning process is individualistic, which is

affected by a persons cognitive ability, physiological state, motivation and emotion, and
interaction between the instructional environment and teacher (Keefe, 1987). This shows
we need to value students individual differences. Before we can teach students according
to their background and strengths, we have to identify students different learning styles.
In literature review, we have yet to locate published inventories for online learning styles.
As different learning styles would affect students learning effects, the present study
recognized the importance to design an online learning styles inventory. Accordingly,
instructors may use the outcome of the inventory to design effective e-learning courses
for students of different learning style preferences. The objectives of the present study
are:
1. Determine learning style categories for online learning assessment
2. Design items of online learning styles
3. Examine the validity and reliability of the online learning style inventory
LITERATURE REVIEW
Features and studies of online learning
With advances in information technology, online courses unrestricted by time and
space are gaining increasing popularity among instructors and learners. Many teachers
have also found learning effective through the use of Web 2.0 technology. For instance,
there are positive reports on using Information and Communication Technology to
support intentional learning (Oshima, Oshima, Yuasa, Konishi, Itoh, & Okada, 2008), the
support of learning using weblogs (Juang, 2008), and the promoting of deep learning
using Wiki (Chua, & Chua, 2008). Thadphoothon (2002) mentioned in their paper that
computer-mediated collaborative learning has the potential to enhance critical thinking
in language learning (p. 1491). Instructors can make good use of the discussion boards
for students to think critically because it would involve writing skills and allow online
class interaction. Students can act as moderators and carry out collaborative learning
through interactive activities.
Features of online learning may be more suitable for some learners, for instance, shy,
independent learners may find online learning more comfortable than traditional
face-to-face learning environment. In addition, as compared with the traditional,
systematic, linear teaching course design, online course arrangement may be more

International Journal of Cyber Society and Education 101

appropriate for students with non-traditional and non-linear learning styles (Illinois
Online Network, 2008). Communication in online learning relies mostly on writing
abilities, and emphasizes on students self-motivation and discipline (lack of teacher
supervision due to unfixed classroom and class time) (Mupingo, Nora, & Yaw, 2006).
Unlike direct contact and interaction in traditional classroom instructions, online
instructors may have no knowledge of learners learning preferences or immediate
responses to the instruction. If learners learning styles were not known beforehand, it
would be difficult for teachers to offer suitable teaching contents and arrangement
according to learners individual learning style preferences.
Definition and categorization of learning styles
In the past, school instruction is focused on course arrangements and teachers
instructional strategies, which is believed to be enough for effective learning. Afterwards,
educationists understand that teaching quality is deeply affected by the students
characteristics, the teachers teaching styles, and teaching environment (Keefe, 1987). A
students characteristics, in terms of learning styles, are referring to an individuals
combination of stable cognitive, affective, and physiological states, thus the students
preferred behavior would be reflected on how they perceive, respond, and interact with
the environment (definition of learning style in NASSP, from Keefe, 1987). Although
cognitive style and learning style were known as synonyms in the past, learning
styles not only include cognitive styles, but also the affective and physiological learning
preferences.
Gregorc (1984) proposed that style reflects an individuals unique, systematic
thoughts and modes of behavior. This is also the behavior model for environmental
adjustment, formed from the interaction among an individuals genes, environment, and
cultural factors. While style is a hypothesized constructive concept, understanding a
persons learning style will be helpful to explain the learning process and further improve
learning effects. If students are exposed to teaching methods inappropriate for their
learning styles, this may result in affective and physiological perceptual problems
(Gregorc, 1979).
Curry (1983) provided the analogy of the structure of learning style as to the peeling
of onions. The core of the onion is an individuals basic personality trait and this trait
measures how this individual accesses and integrates information. The second level is
information-processing, which focuses on the individuals information-processing and

International Journal of Cyber Society and Education 102

cognitive preferences. The third one is social interaction, that is, the effect from
individuals interaction between learning environment and peers. The outer level is more
focused on individual preferences for instruction and environment. Individuals learning
style is stable and difficult to change near the core, while outer levels are prone to be
changed with learning or experience.
The role of learning styles in learning and teaching
Individuals possessing different learning styles indicate respective learning styles
preferences. However, preferences of different learning styles do not lead to different
results in intelligence or academic performance. Dunn (1990) believed learning content
or subject is not the cause of learning failureif an instructor can teach with the styles
the students are good at, anyone can learn effectively. Dunn, Griggs, Olson, Beasley and
Gorman (1995) collected and analyzed 36 studies which adopted Dunns learning style
assessment tools from 1980 to 1990. Results showed students exposed to learning
environments suitable for their learning styles, their average test grades and learning
attitude were three-fourth standard deviation higher than those exposed to the unsuitable
learning environment. In another study, individuals exposed to learning environments
suitable for their learning styles showed improvement in scores (Dunn, 1990). Hence, if
teachers understand the types of learning styles the students possess, and redesign or
adjust the teaching methods to provide learning environment and media appropriate for
students specific learning styles, this may help to improve the learning results for
learners who dislike traditional lecture-based courses. Therefore, diagnosis of students
learning styles provides useful information for educators to arrange or design suitable
teaching methods and teaching environment according to students individual differences
(Keefe, 1987).
Assessment of learning styles
Among the commonly used tests or scales for learning style assessment, the earlier
one is the Group Embedded Figures Test developed by Herman Witkin (1976) and his
colleagues. Participants are asked to locate a specific shape in a complicated figure, in
order to identify if an individual is field independent/analytic or field dependent/global in
processing information. For learning style tests accessing personality, the Myers-Briggs
Type Indicator (Myers, 1978), developed according to Jungs personality theory is able to
identify four personality dimensions, extraversionintroversion, sensationintuition,

International Journal of Cyber Society and Education 103

thinkingfeeling, and judgingperceiving. David Kolb (1976) developed a learning

style scale according to his experience learning model. The model is a four-step cycle:
from the steps of concrete experience, reflective observation, abstract conceptualization,
and then to active verification. Students learning preferences are then categorized as
divergers, assimilators, convergers, and accommodators.
Dunns Learning Style Inventory (for grades 3 through 12) measures 24 factors
(Dunn, Dunn, & Price, 1981), and the Productivity Environmental Preference Survey (for
adult learners) measures 21 factors (Price, Dunn, & Dunn, 1982). Besides different
cognitive or perceptual styles, Dunn et al. believed that individuals may also differ in
environment and social interaction preferences. An individuals learning styles may be
divided into four main categoriesenvironmental (sound, light, temperature, classroom
design), emotion (motivation, persistence, responsibility, need for structure), sociological
(working alone, with others, with an adult), and physical/perceptual preference (visual,
auditory, tactile, kinesthetic, intake, time of day, need for mobility).
Keefe and his associates developed a Learning Style Profile for grades 6-12 students,
measuring 23 factors in totalcognitive skills (analytic, spatial, discrimination,
categorizing, sequential processing, memory), perceptual response (visual, auditory,
emotive), persistence orientation, verbal-spatial, manipulative, study time (early morning,
late morning, afternoon, evening), grouping, posture, mobility, sound, lighting, and
temperature (Keefe & Monk, 1986).
Gregorc Style Delineator (Gregorc, 2004) measures cognitive preferences related to
perception. It is divided into two dimensions: ConcreteAbstract, SequentialRandom,
resulting in four learning styles: Concrete Sequential, Abstract Sequential, Concrete
Random and Abstract Random.
The Index of Learning Styles was developed by Felder and Soloman in 1991 (Felder
& Spurlin, 2005). It is an online instrument and can be submitted and automatically
scored on the Web. It contains 44 items to assess preferences on four dimensions
(active/reflective, sensing/intuitive, visual/verbal, and sequential/global).
Reliability of the learning style tests
The reliability of most of the learning styles inventories is moderate. The Learning
Style Inventory developed by Dunn (1981) and his associates measures 24 learning styles,
with low to moderate reliability coefficients. The authors reported in 1988 that the
coefficient ranged from .55 to .88.

International Journal of Cyber Society and Education 104

The internal consistency of the subscales of the Learning Style Profile ranged
from .47 to .76, and the average coefficient was .61. The authors contributed this low
average reliability to the small number of items that comprise for some sub-scales (Keefe
& Monk, 1986).
The Cronbachs of the Index of Learning Styles for the four dimensions ranged
from .41 to .76; test-retest correlations (4-week, 7-week, and 8-month) were from .50
through .87 (Felder & Spurlin, 2005).
From the review of the literature, almost all of the existing learning style inventories
are adopted in the traditional learning situations. Carnevale pointed out that most studies
indicate that it is difficult to identify the specific learning styles possessed by the online
students, which might have resulted from dispersed learning styles, or assessment tools
adopted inappropriately for the e-learning environment (as cited in Mupingo, Nora, &
Yaw, 2006). Therefore, it is important to develop a test appropriate for learners in online
learning environment, and furthermore, the results of such test would allow the course
designers to develop online courses suitable for learners of multi-learning styles.
METHODOLOGY
According to our research objectives, this section describes the ways we determined
learning style categories for online learning assessment; designed items of online learning
styles; and examined the validity and reliability of the online learning style inventory.
Determine the categories of online learning styles
Online learning features include learning in the unrestricted time and location, but it
requires learners learning self-management. The online learning contents allow multiple
media elements such as audio-visual elements, graphics, textual information, and
hyperlink functions. With reference to learning style categories in related learning style
tests (for example, those from Dunn, Kolb, Keefe, Gregorc, Felder and Soloman), we
identified three categories consisting of 15 factors in the learning styles inventory
developed in the present study:
(i) Perceptual types (this refers to the preference of using particular perceptual sense in
learning): 1. Text: preference for textual information in learning; 2. Visual:
preference for figures and charts information; 3. Auditory: preference for sound and
voice information; 4. Active: preference for touch, hands-on operation, and
self-experience.

International Journal of Cyber Society and Education 105

(ii) Cognitive processing types (this refers to the cognitive tendency for processing
information): 1. Abstract: preference for abstract or conceptual methods for
information process; 2. Concrete: preference for daily experience or concrete
examples; 3. Serial: preference for serial and linear learning; 4. Random: preference
for learning in a non-linear sequence or order; 5. Holistic/Global: preference for
overall understanding of the information; 6. Analytic: preference for detail analysis
of every part of the reading or information.
(iii) Social Interaction and Personality types (this refers to the preferences in social
interaction and personal traits in learning condition): 1. Study alone: preference for
solitary in learning; 2. Study with group: preference for interaction with peers; 3.
Guided: preference for guidance or supervision by an instructor; 4. Persistence: the
tendency to focus in learning in a lengthy amount of time; 5. Observer: prefers
observation instead of involvement in discussion or interaction with others.
Design the items of the inventory
In consideration for a wider range of applications for learners in the future
(especially for learners with no prior online learning experience), most items (i.e., the
questions or statements) in our online learning styles inventory inquires learners learning
preferences in general learning situations. However, two or three items in each learning
style factor measure the online learning situation by using terms like When playing
online games, or When searching for information online to assess participants
preferences in online learning or Internet experiences.
Examine the validity and reliability of the inventory
At the beginning, we developed 168 items. Items were assessed twice by three
scholars in the fields of psychometrics, learning styles, and e-learning. After items were
modified according to their feedback, the content validity of the scales is therefore
assumed to be valid. In addition, the first draft of inventory was administered to five
undergraduate students from different colleges in Taiwan. From their responses and
suggestions, which included the clarification on certain terms and reduction to the
number of items, 105 items were selected for the pilot test.
Pilot test sample. Pilot tests were distributed to undergraduate students of
humanities, science, business, engineering, medicine and agriculture in 10 universities in
different areas of Taiwan. A total of valid 372 questionnaires were returned.

International Journal of Cyber Society and Education 106

Item selection for the main study. As the present inventory was developed to
measure 15 factors of learning styles, these factors may be inter-correlated. When factor
analysis was conducted with all of the data, the result did not show 15 distinct factors as
we hypothesized. Therefore, we decided to conduct factor analysis within each subscale.
Item selection for the main study was based on the results of the factor analysis and
researchers judgment. In addition, for each subscale, at least one online learning item
was selected. Finally, 64 items were selected from 105 items, with four to five items for
each learning style. There were five negative items in the scale.
Reliability analysis. Three reliability analyses were conducted in the present study.
First, internal consistency reliability analysis of the 15 learning styles was conducted for
the 105-item version, with 372 samples; second, a five-week interval test-retest reliability
analysis was conducted with the 64-item version (participants were 35 humanities
undergraduates in a private university); and third, internal consistency reliability analysis
of the 15 learning styles was conducted for the 64-item version (participants were 137
humanities undergraduates in two private universities in North Taiwan, including the first
set of responses from the test-retest sample).
RESULTS AND DISCUSSION
Result of reliability analysis for the 105-item learning styles inventory
Results of the 105-item internal consistency reliability analysis using 372 samples
are shown in Table 1. For five of the 15 learning styles, the Cronbachs was higher
than .60; for another five learning styles, the Cronbachs was between .50 and .60, and
for the rest of the learning styles, it was lower than .50. The low to moderate reliabilities
for the 105- items were similar to the results in the learning style inventories developed
by Dunn et al. (1981) and Keefe et al. (1986).
Result of reliability analysis for the 64-item learning styles inventory
Overall, when we take out the data on 64-items from the 105-items inventory using
the 372 samples, we found those with lower Cronbachs (those less than .50) in the
original 105-item analysis, such as Auditory, Concrete, Serial, were found improved (see
Table 1). The coefficients of the test-retest reliability for the 35 samples on the factors of
Auditory, Concrete, Serial, Random, and Analytic were raised (see Table 1).

International Journal of Cyber Society and Education 107

When our inventory of 105 items was administered in numerous universities, many
respondents expressed that there were too many items to be filled out. Consequently we
tried to reduce as many items as possible in the revised version. Theoretically, with
other factors remaining constant, Cronbachs is lower with fewer items. However, in
our study, when we reduce the items from 105 to 64, the low Cronbachs for the factors
of Auditory, Concrete, Serial, and Random increased. This indicates the
reduction of items improved the reliability of internal consistency.
As for the test-retest analysis, only Active, Holistic/Global was less than .50, the
Pearson correlation coefficients of the other 13 learning styles were more than .50 (see
Table 1).
The results of reliability analysis of the Perceptual, Cognitive-processing, and
Personality learning styles
Reliability analyses (both the internal consistency and test-retest reliability) showed
moderate reliability in the category of Perceptual learning styles. Reliability is lower for
Cognitive-processing learning styles, especially for Serial, Random, and Analytic.
Higher reliability was found in Social Interaction and Personality learning styles,
including Study alone or Study with group, Guided, Persistence, and Observer
(the average was .66 for the 372 samples in the 105-item; average was .63 for the 137
samples in the 64-item, and the average Pearson correlation coefficient was .66 for the 5
weeks test-retest reliability).
Test results on gender differences
Results of the t-test on the 105-item inventory done by 372 samples showed males
had significant lower scores than those of the females on Auditory learning styles, but
higher scores on Abstract, Random, and Analytic learning styles (see Table 2).
From this finding, we can see that the major difference in male and female learning styles
lies in the cognitive-processing category. In other words, more male participants prefer to
process information through abstract or conceptual methods. According to the results in
Philbin, Meier, Huffman, and Boverie study (1995), male adults are primarily abstract
and reflective, and if the learning requires thinking and watching, they learn best; on the
other hand, female adults learn better through watching and by doing. One other study
administrated Gregorc Style Delineator on university Spanish-speaking students in Texas
found males showed more preference with Abstract and Sequential learning styles,

International Journal of Cyber Society and Education 108

and females had stronger preferences on Concrete and Random (Picou, Gatlin-Watts,
& Packer, 1998).
Table 1 Reliability for the 15 factors of learning styles
Sample of 372
participants

Sample of 137
participants

Internal consistency reliability

(Cronbachs )
Factor of learning
style

Sample of 35
participants
5-week test-retest
reliability ( Pearson r)

105-items

64-items a

64-items

64-items

Text

.54

.54

.48

.52

Visual

.54

.54

.56

.77

Auditory

.37

.49

.59

.52

Active

.58

.54

.70

.45

Abstract

.74

.72

.67

.56

Concrete

.46

.72

.72

.51

Serial/Linear

.22

.37

41

.51

Random

.43

.48

.47

.59

Holistic/Global

.53

.52

.44

.49

Analytic

.35

.35

.49

.68

Study alone

.73

.67

.68

.73

Study with group

.72

.65

.76

.77

Guided

.72

.64

.59

.65

Persistence

.52

.45

.51

.53

Observer

.60

.57

.63

.60

The calculation is based on the 64 items selected from the inventory of 105 items.
Note: This table shows the reliability results for the samples of 372 participants (on 105 items and 64 items
respectively), 137 participants (on 64 items), and 35 participants (on 64 items).

International Journal of Cyber Society and Education 109

Furthermore, our study showed that the males do not have to follow serial
procedures or order in learning, but access different chapters or web pages randomly. And
they prefer to analyze each part the information in detail instead of grasping the overall
concept holistically at the beginning of their learning. On the other hand, female
participants have higher scores in the Auditory factor than males. It appears that female
students prefer voice and sound for receiving and processing information in learning
when compared to male students. This is consistent with Honigsfeld and Dunn finding
that females in many nations have been more auditory than their male counterparts
(2006).
As for the results from the 64-item inventory done by 137 samples, the outcome
showed significant differences between male and female learning styles in Abstract and
Study with group (see Table 2). That is, males prefer to process information in abstract
or conceptual methods more than females in learning and this is consistent with the result
of the 372 sample. In addition, male students prefer to interact and learn with peer more
than female students during their learning, and a similar finding was found in Honigsfeld
and Dunn study (2003). Results from these two samples were not entirely consistent,
which may be caused by sample differences (i.e., this sample was taken only from the
humanities college), and differences in the number of items for both samples.
Test results on college differences
To examine learning style differences among different colleges, the present study
divided the 105-item inventory done by 372 samples into 4 main college categories:
Humanities College (n = 59, 15.9%), Business and Management College (n = 115,
30.9%), Science College (n = 40, 10.8%), and Engineering College (n = 157, 42.4%).
Analysis of variance (ANOVA) was conducted, and the Scheff post hoc test was
applied for significant ANOVA results. The ANOVA results revealed significant
differences among the colleges for the following factors of learning style: Visual
(F(3,357)=3.301, p<.05), Auditory (F(3,358)=3.207, p<.05), Abstract
(F(3,361)=8.880, p<.001), Analytic (F(3,358)=3.352, p<.05), and Guided
(F(3,365)=7.043, p<.001) (see Table 3).

International Journal of Cyber Society and Education 110

Table 2 Results of gender difference on learning styles

Sample of 372 participants
(105-items)
Factor of learning
style

Male
M

Sample of 137 participants

(64-items)

Female
M

Male

Female

Text

3.28

3.26

.36

.72

3.33

3.32

.159

.88

Visual

3.60

3.55

1.07

.29

3.63

3.79

-1.59

.12

Auditory

3.50

3.60

-2.13

.03

3.69

3.81

-1.11

.27

Active

3.73

3.69

.80

.42

3.95

3.79

1.27

.21

Abstract

3.42

3.29

2.28

.02

3.32

3.02

2.57

.01

Concrete

3.63

3.68

-1.13

.26

4.03

4.13

-1.07

.29

Serial/Linear

3.23

3.24

-.17

.87

3.52

3.60

-.90

.37

Random

3.41

3.28

2.44

.02

3.56

3.39

1.56

.12

Holistic/Global

3.48

3.49

-.18

.86

3.49

3.49

.01

.99

Analytic

3.27

3.17

2.18

.03

3.22

3.10

1.17

.25

Study alone

3.16

3.21

-.74

.46

3.23

3.45

-1.58

.12

Study with group

3.44

3.36

1.16

.25

3.72

3.44

2.15

.03

Guided

3.36

3.34

.36

.72

3.44

3.59

-1.34

.18

Persistence

2.85

2.95

-1.76

.08

2.76

2.92

-1.22

.22

Observer

3.14

3.09

.74

.46

3.15

3.17

-.13

.89

p <.05
Note: a M= mean, b t= t-score, c p= p-value

Results of the Scheff post hoc test (the most conservative method, Huck, 2000)
revealed only four learning styles factors showed significant differences between groups
(Auditory, Abstract, Analytic, and Guided): First, the Science College sample
yielded significant lower scores in Auditory learning style than the
Business/Management and Engineering College samples; second and third, the
Engineering College samples showed significant higher Abstract learning style scores

International Journal of Cyber Society and Education 111

than Humanities and Business/Management College samples, and significant higher

Analytic learning style scores than the Business/Management College samples; and
lastly, the Business/Management College samples obtained the highest scores in
Guided learning style among the college samples.
Table 3 Results of college difference on learning styles
Factor of learning
style
Text

Humanities

Business/
Management

Science

Engineering
Fa

M
3.25

3.22

3.35

3.28

.78

.51

Visual

3.48

3.52

3.67

3.63

3.30

Auditory

3.55

3.57

3.35

3.58

3.21

Active

3.68

3.76

3.66

3.70

.65
8.88

.02
.02
.58

Abstract

3.15

3.23

3.46

3.50

Concrete

3.69

3.68

3.55

3.65

1.34

.26

Serial/Linear

3.22

3.26

3.19

3.24

.41

.75

Random

3.31

3.32

3.38

3.37

.37

.78

Holistic/Global

3.47

3.49

3.34

3.54

1.95

.12

Analytic

3.17

3.14

3.30

3.29

3.35

Study alone

3.30

3.12

3.27

3.16

1.15

.33

Study with group

3.32

3.45

3.20

3.45

2.45

.06

Guided

3.20

3.52

3.21

3.31

7.04

.00

Persistence

3.01

2.89

2.86

2.88

.90

.44

Observer

3.05

3.09

3.16

3.15

.77

.51

.00

.02

p <.05, p <.001
Note: aF= F score

These results indicated that Science students do not prefer to receive information
through auditory than Engineering and Humanities students. Engineering students prefer
to think more abstractly and conceptually than Humanities, Business and Management

International Journal of Cyber Society and Education 112

students, and emphasize more in analyzing detail parts of each section. This may be
related to the study of more abstract theories in physics and chemistry for engineering
students, hence requiring the need to verify experimental procedures or formula. In
addition, Business and Management students are in favor of being guided or supervised
by an instructor in learning than students in the other colleges. This may be due to the
emphasis of team work and supervision in business and management field. A study in
Singapore by Yuen and Lee (1994) revealed the similar results: The students who were in
social and humanities fields showed more preferences in the category of Diverger, that
is, they were concrete experiencer and reflective observer; as for the science students,
they tended to be Assimilators, which means they preferred abstract thinking and
conceptualized their understanding.
CONCLUSION AND RECOMMENDATIONS
In the present study, we made an attempt to develop an online learning styles
inventory. We referenced important factors in learning style inventories in literature, and
with consideration of e-learning features, 15 factors (105 items in total) related to
e-learning styles were selected from the original 168 items. Results of the internal
consistency from the pilot sample showed higher reliability, close to .60 or .70, in
Active, Abstract, Concrete, Study alone, Study with group, Guided, and
Observer styles. However, lower reliability, mostly less than .50, was shown in
Auditory, Serial, Random, and Holistic/Global styles. After factor analysis and
researchers judgment, 105 items were reduced to 64 items (with 4 to 5 items in each
factor). A new sample test and a 5-week test-retest reliability analysis were conducted.
Learning style factors with lower reliability value in the 105-item scale showed increase
in this shortened scale.
Reliability study of the present inventory of learning style preferences is at the
beginning phase. The reliability coefficients of the sub-scales in this inventory range from
low to moderate. In order to improve the reliability of the scale, continuous empirical
study to verify the reliability and consistent item revision are required.
Results of learning style differences in the pilot test (105-item filled out by 372
participants) and the second undergraduate samples (64-item filled out by 137
participants) showed significant differences in gender and college in some learning styles.
As the present inventory is still at the development stage and the number of samples
accumulated is small, hence, more studies need to be conducted to establish norms for

International Journal of Cyber Society and Education 113

different types of participants in the future.

The present inventory has 15 factors (in three large categories) which are related to
the online learning situation. It can help instructors or online course designers understand
students preferred learning styles in the dimension of the perceptual types, cognitive
processing types, and Social Interaction and Personality types, as a reference to design
suitable courses and adopt appropriate teaching methods. For example, for students who
preferred to learn through auditory materials, the online course should include various
types of audio files such as teachers lectures in audio formats and guest speakers speech
contents; for learners who prefer active learning, instructors should offer more materials
such as 3D animations and/or online puzzles; for those who like to study with peers,
teachers should increase the opportunities for synchronous discussion (e.g, via MSN
Messenger or JoinNet videoconference programs); for the ones who need to be monitored
constantly in their learning process, the course designers should include checklists and
assessments to guide them. Furthermore, the results obtained from the inventory can also
be used to compare with students learning performances in order to improve the contents
of the online course. However, it should not be used exclusively for student selection or
categorization. In order to better stabilize and improve this inventory, more studies are
needed to examine its validity and reliability.
The researchers have been collecting data on the relationships between students
learning styles and their grades, such results can be the evidence of criterion-related
validity. In addition, future studies on assessing learning styles of students across
different cultures may also be explored. These findings will be valuable to teachers who
offer online courses to students of diverse backgrounds.
REFERENCES
Chua, G-K., & Chua, G-B. (2008). Organising collaborative learning spaces for
knowledge construction: Deep learning and online behavior. In T. W. Chan, G.
Biswas, F. Chen, S. Chen, C. Chou, M. Jacobson, et al. (Eds.), Proceedings of the
16 th International Conference on Computers in Education (ICCE2008) (pp.
285-289). Taipei, Taiwan: Asia-Pacific Society for Computers in Education.
Curry, L. (1983, April). An organization of learning styles theory and construct. Paper
presented at the annual meeting of the American Educational Research Association,
Montreal, Canada.

International Journal of Cyber Society and Education 114

Dunn, R., Dunn, K., & Price, G. E. (1981). Learning Style Inventory. Lawrence, KS:
Price System.
Dunn, R., Griggs, S. A., Olson, J., Beasley, M., & Gorman, B. S. (1995). A meta-analysis
validation of the Dunn and Dunn Model of Learning-Style Preferences. The Journal
of Educational Research, 88(6), 353-362.
Dunn, R. (1990). Rita Dunn answers questions on learning styles. Educational
Leadership, 48(2), 15-19.
Felder, R. M., & Spurlin, J. (2005). Applications, reliability, and validity of the Index of
Learning Styles. International Journal of Engineering Education, 21(1),103-112.
Gregorc, A. F. (1979). Learning/Teaching styles: Potent forces behind them. Educational
Leadership, 36, 234-236.
Gregorc, A. F. (1984). Style as a symptom: A phenomenological perspective. Theory into
Practice, 23(1), 51-55.
Gregorc, A. F. (2004). Gregorc Style Delineator (3rd ed.). Columbia, CT: Gregorc
Associates, Inc.
Honigsfeld, A., & Dunn, R. (2003). High school male and female learning-style
similarities and differences in diverse nations. Journal of Educational Research,
96(4), 195-206.
Honigsfeld, A., & Dunn, R. (2006). Learning-style characteristics of adult learners. The
Delta Kappa Gamma Bulletin, 72(2), 14-17, 31.
Huck, S. W. (2000). Reading statistics and research (3rd ed.). New York: Longman.
Illinois Online Network (2008). Learning styles and the online environment. Retrieved
March
18,
2008,
from
http://www.ion.uillinois.edu/resources/tutorials/id/learningStyles.asp#top
Juang, Y-R. (2008). Learning by blogging: Warm-up and review lessons to facilitate
knowledge building in classrooms. In T. W. Chan, G. Biswas, F. Chen, S. Chen, C.
Chou, M. Jacobson, et al. (Eds.), Proceedings of the 16th International Conference
on Computers in Education (ICCE2008) (pp. 279-283). Taipei, Taiwan: Asia-Pacific
Society for Computers in Education.
Keefe, J. W. (1987). Learning style theory and practice. Reston VA: National
Association of Secondary School Principals.
Keefe, J. W., & Monk, J. S. (1986). Learning styles profile examiners manual. Reston,
VA: National Association of Secondary School Principals.
Kolb, D. (1976). Learning Styles Inventory. Boston: McBer and Co.

International Journal of Cyber Society and Education 115

Myers, I. (1978). Myers-Briggs Type Indicators. Palo Alto, CA: Counseling

Psychologists Press.
Mupingo, D. M., Nora, R. T., & Yaw, D. C. (2006). The learning styles, expectations,
and needs of online students. College Teaching, 54(1), 185-189.
Oshima, J., Oshima, R., Yuasa, K., Konishi, T., Itoh, Y., & Okada, Y. (2008). Facilitation
of intentional learning and formative assessments in large-scale university courses
with ICT. In T. W. Chan, G. Biswas, F. Chen, S. Chen, C. Chou, M. Jacobson, et al.
(Eds.), Proceedings of the 16th International Conference on Computers in Education
(ICCE2008) (pp. 275-278). Taipei, Taiwan: Asia-Pacific Society for Computers in
Education.
Philbin, M., Meier, E. Huffman, S., & Boverie, P. (1995). A survey of gender and
learning styles. Sex Roles, 32(7/8), 485-494.
Picou, A., Gatlin-Watts, R., & Packer, J. (1998). A Test for learning style differences for
the U.S. border population. Texas Papers in Foreign Language Education, 3 (2),
105-116.
Price, G. E., Dunn, R., & Dunn, K.(1982). Productivity Environmental Preference Survey.
Lawrence, KS: Price System.
Thadphoothon, J. (2002). Enhancing Critical Thinking in Language Learning through
Computer Mediated Collaborative Learning: A Preliminary Investigation. In
Kinshuk, R. Lewis, K. Akahori, R. Kemp, T. Okamoto, L. Henderson, et al.
Proceedings of the 2002 International Conference on Computers in Education (pp.
1490-1491). Auckland, New Zealand: IEEE Computer Society.
Witkin, H. A. (1976). Cognitive style in academic performance and in teacher-student
relationships. In S. Messick & Associates (Eds.), Individuality in learning. San
Francisco: Jossey-Bass.
Yuen, C. & Lee S. N. (1994). Learning styles and their Implications for cross-cultural
management in Singapore. Journal of Social Psychology, 134(5), 593-600.
ACKNOWLEDGEMENT
This study is supported by National Science Council of Taiwan, R.O.C. under
contract number NSC 95-2520-S-030-002-MY2.

International Journal of Cyber Society and Education 116