This article was downloaded by: [66.31.30.

1]
On: 02 January 2015, At: 19:41
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Journal of Geography
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/rjog20

Components of Spatial Thinking: Evidence from a

Spatial Thinking Ability Test
a b
Jongwon Lee & Robert Bednarz
a
Ewha Womans University , Seoul , South Korea
b
Texas A & M University , College Station , Texas , USA
Published online: 27 Dec 2011.

To cite this article: Jongwon Lee & Robert Bednarz (2012) Components of Spatial Thinking: Evidence from a Spatial Thinking
Ability Test, Journal of Geography, 111:1, 15-26, DOI: 10.1080/00221341.2011.583262

To link to this article: http://dx.doi.org/10.1080/00221341.2011.583262

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Components of Spatial Thinking: Evidence from a Spatial
Thinking Ability Test
Jongwon Lee and Robert Bednarz

ABSTRACT
This article introduces the development and
validation of the spatial thinking ability INTRODUCTION
test (STAT). The STAT consists of sixteen Spatial thinking has been actively investigated during the last decade, especially
multiple-choice questions of eight types.
The STAT was validated by administering it with respect to its relationship to geospatial technologies and its relevance
to a sample of 532 junior high, high school, to problem solving in everyday life, the workplace, and science (Albert and
and university students. Factor analysis Golledge 1999; Battersby, Golledge, and Marsh 2006; Bednarz 2004; Golledge 2002;
using principal components extraction was Marsh, Golledge, and Battersby 2007). However, long before researchers began
applied to identify underlying spatial
thinking components and to evaluate to focus on spatial thinking, psychologists and others sought to identify and
the construct validity of the STAT. measure spatial ability. Spatial ability—typically defined as spatial perception,
Spatial components identified through visualization, and orientation—is seen as a narrower concept than spatial thinking
factor analysis only partly coincided with (Committee on Support for Thinking Spatially 2006). It is beyond the scope of
spatial concepts used to develop the
questions that compose the STAT and this article to provide a comprehensive review of the literature concerning the
with the components of spatial thinking differences and distinctions between spatial ability, spatial reasoning, spatial
Downloaded by [66.31.30.1] at 19:41 02 January 2015

hypothesized by other researchers. cognition, spatial concepts, spatial intelligence, and environmental cognition.
Learning to Think Spatially, published by the National Research Council, while
Key Words: spatial thinking, spatial recognizing that no clear consensus as yet exists concerning spatial thinking,
thinking ability test (STAT), factor analysis provided a significant step toward understanding its nature and its importance
in the school curriculum. The Committee (26) saw spatial ability as “a trait that a
person has and as a way of characterizing a person’s ability to perform mentally
such operations as rotation, perspective change, and so forth. The concept
derives in part from the psychometric tradition of intelligence measurement
and testing . . . ” The Committee viewed spatial thinking, on the other hand, as a
constructive amalgam of three mutually reinforcing components: the concept of
space, tools of representation, and processes of reasoning. In order for individuals
to conceptualize space, understand representations, and reason spatially, they
must possess the appropriate spatial skills (Committee on Support for Thinking
Spatially 2006).
The Committee (2006) also recognized the educational value of spatial thinking,
arguing that it can be taught and learned; thus spatial thinking should be an
important part of the educational curriculum at all levels. The Committee further
suggested that GIS and other geospatial technologies can play a powerful role in
promoting spatial thinking. In fact, many studies have pointed to the advantage
of integrating GIS into the classroom (e.g., Allen 2007; DeMers and Vincent 2007;
Doering and Veletsianos 2007; Milson and Earle 2007; Patterson, Reeve, and Page
Dr. Jongwon Lee is an assistant professor of 2003) and have shown explicit links between GIS learning and students’ spatial
social studies education (geography education
major) at Ewha Womans University, Seoul, thinking skills (Kerski 2008; Lee and Bednarz 2009; Schultz, Kerski, and Patterson
South Korea. His research interests are in the 2008).
areas of geography education: spatial cognition However, researchers have also argued that “to be most effective, GIS teaching
and skills, especially the role of geospatial and curriculum development strategies should begin with an assessment of
technology in spatial thinking.
student understanding of spatial relationships. . . ” (Wigglesworth 2003, 282),
emphasizing the importance of establishing viable spatial thinking assessment
Robert Bednarz is professor of geography, Texas
A & M University, College Station, Texas, USA, based on a scientifically rigorous definition (Eliot and Czarnolewski 2007).
and North American Commissioning Editor of Unfortunately, such a standardized measure of essential knowledge and skills
the Journal of Geography in Higher Education. does not exist. In fact, the Committee stated explicitly that “[t]here are neither
His recent research has focused on spatial think-
ing, the impact of using geospatial technologies
content standards nor valid and reliable assessments for spatial thinking”
on spatial thinking skills, assessment of spa- (Committee on the Support for Thinking Spatially 2006, 232).
tial thinking skills, and the implementation of This article begins with a brief discussion of concepts of spatial thinking
geospatial technologies in science and geoscience skills and the instruments available to measure them. Next, the article presents
curricula.
the development and validation procedures of the spatial thinking ability test

Journal of Geography 111: 15–26

C 2012 National Council for Geographic Education 15
 Jongwon Lee and Robert Bednarz

(STAT) that is modeled after the spatial skills test (Lee organization of the world economy” (Bednarz n.d.). The
and Bednarz 2009). Data are presented that support the concept or object investigated through the third context is
validity and reliability of STAT based upon a field test not necessarily spatial but can be spatialized by time-space
of 532 junior high, high school, and university students. coordination. For example, written linguistic symbols are
The differences in the performance of these three levels spatially defined and spatially arranged, and readers must
of students are explored and tested for significance using establish word order so that sentence and passage meaning
ANOVA. In addition, factor analysis is applied to identify can be determined. Patterns in complex numerical data can
underlying spatial thinking components, to determine if often be revealed and best understood by portraying the
the identified components support the structure of spatial information graphically.
thinking proposed by other researchers, and to evaluate the Although Learning to Think Spatially (Committee on the
construct validity of the STAT. Support for Thinking Spatially 2006) provides multicontex-
tual and interdisciplinary definitions of spatial thinking, it
has been criticized for its lack of a conceptual framework,
CONCEPTS OF SPATIAL THINKING an essential prerequisite for development of assessment
Because the term “spatial thinking” has been used in both tools (e.g., Gersmehl and Gersmehl 2006). Previous research
nonacademic and academic areas extensively, a variety of is not devoid of conceptual frameworks, however, as
definitions exist (Committee on the Support for Thinking Gersmehl and Gersmehl (2006, 2007) and Golledge and
Spatially 2006; Eliot and Czarnolewski 2007; Gersmehl 2005; others (Battersby, Golledge, and Marsh 2006; Golledge
Gersmehl and Gersmehl 2006; 2007; Golledge and Stimson 1995, 2002; Golledge and Stimson 1997; Golledge, Marsh,
1997; Harris 1981; Marsh, Golledge, and Battersby 2007; and Battersby 2008; Marsh, Golledge, and Battersby 2007)
Downloaded by [66.31.30.1] at 19:41 02 January 2015

Montello et al. 1999). In addition, substantial disagreement have proposed hierarchies of spatial thinking skills and
continues to occur about the scale (from tabletop scale concepts. In a study to specify a taxonomy of spatial
to geographic scale) and dimensions (thinking in, about, thinking, Gersmehl and Gersmehl (2006, 2007) defined
and with space) of spatial thinking, about the nature of spatial thinking as skills that geographers use to analyze the
cognitive processes involved, about the number of major spatial relationships in the world. They identified thirteen
components, and, as noted in the introduction, about the modes of spatial thinking: defining a location, describing
relationship, if any, between spatial ability and spatial conditions (the geographic concept of site), tracing spatial
thinking. connections (situation), making a spatial comparison, infer-
A few studies provide valuable input for the develop- ring a spatial aura (influence), delimiting a region, fitting a
ment of spatial thinking assessments. These studies suggest place into a spatial hierarchy, graphing a spatial transition,
a series of spatial thinking concepts and describe differences identifying a spatial analog, discerning spatial patterns,
between expert and novice performance in spatial thinking. assessing a spatial association, designing and using a spatial
Some of the most useful studies include Learning to Think model, and mapping spatial exceptions. They argued that
Spatially (Committee on the Support for Thinking Spatially brain research suggests that these modes of spatial thinking
2006), Gersmehl’s (2005) spatial thinking taxonomy, and have distinct or independent neurological foundations.
Golledge and others’ categorization of geospatial concepts They offered no empirical evidence or other rigorous
(Battersby, Golledge, and Marsh 2006; Golledge 2002; assessment to support their hypothesis that the modes they
Golledge and Stimson 1997; Golledge, Marsh, and Battersby identified are independent, however.
2008; Marsh, Golledge, and Battersby 2007). A hierarchical set of spatial thinking concepts was pro-
Learning to Think Spatially (Committee on the Support for posed by Golledge and his colleagues (2008). This sequence
Thinking Spatially 2006) introduces three spatial contexts: progresses from four basic spatial concepts (or primitives)
life space (cognition in space), physical space (cognition to more complex and abstract concepts through five differ-
about space), and intellectual space (cognition with space). ent levels as follows (Golledge, Marsh, and Battersby 2008,
The first of these involves thinking about the world in which 91–92): (1) primitive level (identity, location, magnitude,
we live. It often includes way-finding and navigation in space-time); (2) simple level (arrangements, distribution,
the real, geographic world. It also includes other everyday line, shape, boundary, distance, reference frame, sequence);
activities including assembling it furniture by following (3) difficult level (adjacency, angle, classification, coordi-
instructions, packing the trunk of a car to maximize carrying nate, grid pattern, polygon); (4) complicated level (buffer,
capacity, etc. The second context, cognition about space, connectivity, gradient, profile, representation, scale); and
focuses on “a scientific understanding of nature, structure, (5) complex level (area association, interpolations, map
and function of phenomena that range from the microscopic projection, subjective space, virtual reality). The spatial
to the astronomical scales” (30). It is useful in explaining thinking skills suggested by the Gersmehls (2006, 2007)
the structure of the atom or DNA, the movement and and geospatial concepts proposed by Golledge, Marsh,
arrangement of the elements of the solar system, etc. and Battersby (2008) share the common context of the
Other examples include “shapes and structures of urban geographic scale. However, whereas Gersmehls’ concepts
areas, the diffusion of cultures and agriculture, or the are related to geographic analysis, the geospatial concepts

16
 Components of Spatial Thinking: Evidence from a Spatial Thinking Ability Test

Golledge, Marsh, and Battersby

Table 1. Core concepts of spatial thinking suggested by Gersmehl and Gersmehl (2007), (2008), and Janelle and Goodchild
Golledge, Marsh, and Battersby (2008), and Janelle and Goodchild (2009). (2009) are summarized in Table 1.
Although the terms and number of
core concepts that they used are
Gersmehl and Golledge et al. Janelle and Goodchild
different, it is not difficult to find
Gersmehl (2007) (2008) (2009)
similarity among them.
Condition Identity Objects and Fields
Location Location Location MEASUREMENT OF SPATIAL
Connection Connectivity Network THINKING ABILITIES
Distance Distance
A variety of psychometric tests
Scale Scale
(Clements et al. 1997; Dean and
Comparison Pattern Matching
Aura Buffer Morris 2003; Hall-Wallace and
Region Adjacency, Classification Neighborhood and McAuliffe 2002) have been widely
Region used to measure individuals’ spa-
Hierarchy tial abilities, especially in psycho-
Transition Gradient, Profile logical research. However, psycho-
Analogy metric measures are limited to the
Coordinate assessment of psychologically and
Pattern Pattern, Arrangement, Distribution, narrowly defined spatial abilities
Downloaded by [66.31.30.1] at 19:41 02 January 2015

Order, Sequence rather than spatial thinking as

Spatial Association Spatial Association, Spatial Dependence,
defined by the Committee (2006)
Overlay/Dissolve, Interpolation Spatial Heterogeneity
(Hegarty et al. 2002; Lee and Bed-
Projection, Transformation
narz 2009). Consistent with this
view, Eliot and Czarnolewski (2007,
identified by Golledge and his colleagues are intended 362) argued that “researchers need
primarily to address the functions of GIS. to go beyond the limits of existing spatial tests and
The present investigation, using experiments that began consider the possibility that spatial intelligence is a more
in 2006, could not benefit from the most recent work encompassing construct of human activities . . . .”
concerning hierarchical geospatial ontology. However it Self-assessment questionnaires are believed to assess
did incorporate key spatial thinking concepts from several broader aspects of spatial thinking (Hegarty et al. 2002),
studies conducted by Golledge (1992, 1995, 2002). Along and there are a few examples of these available on the Web
with Gersmehls’ spatial thinking taxonomy, Golledge’s list (e.g., Golledge 2000, 2001). A typical question from these
of geographic thinking elements presented in 2002 guided instruments might ask people to rate on a five- or seven-
the development of the spatial thinking ability test on which point scale a statement such as, “When traveling, I take
this study is based. The following list from Golledge’s 2002 shortcuts as frequently as possible.” Although researchers
study specifies the spatial thinking elements he thought have found that self-report measures are capable of assess-
were important and illustrates the ideas and concepts his ing spatial skills on both the small (or pictorial) and large (or
work shares with Gersmehls’ (2005): environmental) scales (Hegarty et al. 2002) and are useful
in assessing individuals’ spatial behaviors in everyday life,
Comprehending spatial association (pos- they are more appropriate for classifying types of spatial
itive and negative); comprehending spa- behavior than determining levels of spatial ability. Another
tial classification (regionalization); com- shortcoming of subjective self-report measures is that often
prehending spatial change and spatial the results from different instruments are incomparable.
spread (spatial diffusion); comprehending There have been some important attempts to measure
non-spatial and spatial hierarchy; com- specific aspects of spatial thinking skills (e.g., Albert
prehending spatial shapes and patterns; and Golledge 1999; Battersby, Golledge, and Marsh 2006;
comprehending locations and places; com- Gilmartin and Patton 1984; Golledge 1992; Kerski 2000;
prehending integration of geographic fea- Lloyd and Bunch 2003). For example, Golledge (1992) inves-
tures represented as points, networks, and tigated how completely people understand spatial concepts
regions; comprehending spatial closure (in- such as “nearest neighbor” using a map-based laboratory
terpolation); and recognizing spatial form. experiment. Battersby, Golledge, and Marsh (2006) devised
(Golledge 2002, 4–6) a task assessing individuals’ understanding and ability
to apply one of the most essential GIS functions—map
The core concepts of spatial thinking from three recent im- overlay. In that study, participants were provided two maps
portant sources including Gersmehl and Gersmehl (2007), of the same area and asked to derive conclusions about

17
 Jongwon Lee and Robert Bednarz

the spatial relationships. In order to answer the questions isolines based on point data, and differentiating among
properly, participants had to combine two thematic layers spatial data types.
of information and perform logical functions (i.e., Boolean The initial motivation to revise and augment the original
logic). In a similar study, Albert and Golledge (1999) used spatial skills test was to measure students’ mastery of the
a simplified set of thematic layers to evaluate how well GIS content and skills contained in the Association of American
users could select appropriate map layers and operations Geographers’ Teachers’ Guide to Modern Geography (TGMG)
and visually verify map overlay processes to achieve a project materials. The primary aim of the TGMG, funded
specific result. by the U.S. Department of Education, is to improve the
Another type of spatial task used to assess individuals’ preparation and ability of geography teachers to incorpo-
spatial ability includes map-reading skills such as following rate spatial thinking skills into their classes. The TGMG
directions, judging distances, comprehending geographic project produced a variety of print and digital materials
characteristics, and recognizing patterns (Carswell 1971; for preservice and in-service teacher preparation programs,
Gilmartin and Patton 1984). Map-reading skills devised by for example, a multimedia CD with animated instructional
Gilmartin and Patton (1984) provided students with repre- units that deal with the analytical skills specified in the
sentations of a country’s population distribution, topogra- National Geography Standards, such as measuring direction,
phy, and climate that they then used to answer multiple- distance, slope, and density; analyzing map patterns and
choice questions such as “Which of the country’s three making rigorous map comparisons; formulating and testing
major cities has the largest population?” Also included hypotheses; identifying exceptions to patterns predicted
in the same study was a road map-reading task to assess by hypotheses; and buffering, overlaying, windowing, and
abilities such as distance estimation, route comparison (e.g., other methods of spatial analysis. The spatial thinking
Downloaded by [66.31.30.1] at 19:41 02 January 2015

visually compare two straight-line distances and judge ability test (STAT) was designed to assess individuals’
which is shorter), and pattern recognition (e.g., choose growth in spatial thinking skills and to help determine
which of four generalized diagrams best represents the the effectiveness of the TGMG materials in promoting
overall road pattern in the study area). Finally, Kerski the spatial thinking skills of teachers. The revised and
(2000) created a task that assessed both the spatial concept, expanded spatial skills test also provided a data set that
“best location,” and map reading simultaneously. He asked can be used to provide a preliminary assessment of the
students to analyze geographic information and select the reliability and validity of the previously noted spatial
best location for a fast food restaurant in a hypothetical area thinking conceptualizations proposed by other researchers.
based on a given set of variables including traffic volume, The initial step in the construction of the STAT was the
existing fast food locations, locations of high schools, and delineation of the assessment objective and the description
annual median income. of the test contents to be measured. Two sets of spatial
thinking concepts were analyzed and combined and served
to inform the development of STAT. The first set of concepts
DEVELOPMENT OF THE SPATIAL THINKING ABILITY was identified by Gersmehl (2005) whose ideas served as
TEST (STAT) the theoretical foundation of the TGMG project. The second
One goal of the present study was to develop a standard- set was comprised of Golledge’s (2002) list of spatial think-
ized test of spatial thinking abilities (the spatial thinking ing skills, which played a key role in the development of
ability test (STAT)) that integrates geography content the original spatial thinking ability test. Golledge’s concepts
knowledge and spatial skills. Currently no standardized were especially useful because they were detailed enough
instrument for assessing the set of spatial thinking skills to develop test items, potentially leading to improvement
discussed previously exists. In addition, the reviewed of test content validity. Additionally, both lists share some
studies using questionnaires or other measures to assess common concepts and features as noted previously.
spatial skills often ignored issues of reliability and validity. Each test item was designed to measure one or two
The current study extends the authors’ research (Lee and components of spatial thinking identified by one or both of
Bednarz 2009) that developed and deployed spatial skills these two studies. The aspects of spatial thinking abilities
tests (SST) to measure changes in students’ spatial skills covered by STAT include: (1) comprehending orientation
after they completed GIS coursework. That research found and direction; (2) comparing map information to graphic
a significant relationship between the completion of one or information; (3) choosing the best location based on several
more geospatial technology courses and students’ scores on spatial factors; (4) imagining a slope profile based on
the spatial skills test. The components of spatial relations a topographic map; (5) correlating spatially distributed
as defined by Golledge and Stimson (1997) provided guide- phenomena; (6) mentally visualizing 3-D images based
lines for developing test items. The spatial skills tests consist on 2-D information; (7) overlaying and dissolving maps;
of a set of multiple-choice questions and performance tasks and (8) comprehending geographic features represented as
that were designed to evaluate students’ skills including point, line, or polygon (see Table 2).
overlaying and dissolving a map, reading a topographic During the development of STAT, we focused on a
map, evaluating several factors to find the best location, central problem related to test construction: how to
recognizing spatially correlated phenomena, constructing ensure practicability while at the same time providing

18
 Components of Spatial Thinking: Evidence from a Spatial Thinking Ability Test

maximum comprehensibility of
Table 2. Description of question types and spatial thinking components to measure. spatial thinking concepts. A num-
ber of other factors were also
considered in the design of the
Type Item Spatial Thinking STAT. These factors included (1)
(Item Number) Description Components to Measure
cognitive process (i.e., maximizing
I (#1, #2) In order to solve item #1 and #2, Item #1 and #2 evaluate the trait spatial processes and minimizing
participants should visually of “comprehending orientation verbal processes); (2) psychometric
navigate road maps using and direction (e.g., rationale; (3) mode of represen-
verbal information including forward-backward; left-right; tation (text, picture, graph, map,
participant’s current location, up-down; back-front; color versus black and white, etc.);
directions to destination, street horizontal-vertical; and (4) practical constraints (e.g.,
information, etc. (See Fig. 1) north/south/east/west)” amount of time required to com-
(Golledge 2002).
plete the test).
II (#3) In order to solve item #3, Item #3 assesses the trait of
The current version of the test
participants should recognize “discerning spatial patterns”
map patterns and represent (Gersmehl 2005) and “graphing is fourteen pages long and has
them in graphic form. a spatial transition” (Gersmehl two equivalent forms (one that
2005). can be used for a pretest and
III (#4) In order to solve item #4, The basic rationale behind item #4 one for a post-test) allowing for
participants should select an is to assess the trait the evaluation of changes in spa-
Downloaded by [66.31.30.1] at 19:41 02 January 2015

ideal location for a fictitious “comprehending overlay and tial thinking skills over a period
facility based on multiple pieces dissolve” (Golledge 2002) and of time. The pre- and post-tests
of spatial information such as “inferring a spatial aura were composed of slightly different
land use, elevation, population (influence)” (Gersmehl 2005). questions covering the same spatial
density, etc.
thinking skills. Each form, contain-
IV (#5) In order to solve item #5, In solving item #5, participants
ing sixteen multiple-choice ques-
participants should create a deal with several cognitive traits
profile of topography along a including “recognizing spatial tions, consists of eight different
proposed line on a contour map. form (such as cross-sections to types of questions (Table 2). Figure
In addition, the participants three-dimensional block 1a and 1b contain a sample of items
need to properly orient diagrams or image)” (Golledge from the STAT. We also constructed
themselves in situ. 2002), “being able to transform a three-item questionnaire to col-
perceptions, representations lect information about the subject’s
and images from one dimension gender, academic major (geogra-
to another and the reverse” phy major or not), and amount of
(Golledge 2002) and “graphing geospatial coursework completed
a spatial transition” (Gersmehl
(e.g., GIS and cartography).
2005).
Formal and informal review
V (#6, #7) In order to solve item #6, Item #6 and #7 evaluate the trait
participants should identify “comprehending spatial of STAT took place in a vari-
spatial correlations between association (positive and ety of venues, mostly conducted
sets of maps. Additionally, item negative)” (Golledge 2002), as part of the evaluation plan
#7 asks participants to display “making a spatial comparison” of the TGMG project. After a
the identified spatial relationship (Gersmehl 2005), and draft of STAT was completed, all
in a graphic form. (See Fig. 1) “assessing a spatial items were carefully reviewed by
association” (Gersmehl 2005). a team of experts consisting pri-
Item #7 additionally assesses marily of the TGMG project team
the trait of “graphing a spatial and steering committee members.
transition” (Gersmehl 2005).
The team included two individ-
VI (#8) In order to solve item #8, Item #8 assesses the trait of
uals who teach geographic edu-
participants need to mentally “being able to transform
visualize a 3-D image based on perceptions, representations cation courses for undergraduate
2-D information. (See Fig. 1) and images from one dimension and graduate students; these mem-
to another and the reverse” bers conducted an informal review
(Golledge 2002). and then pilot-tested the instru-
VII (#9, #10, In order to solve item #9, #10, #11, Item #9, #10, #11, and #12 ment with their students. Twenty-
#11, #12) and #12, participants should correspond to the trait seven undergraduate students par-
visually verify a map overlay “overlaying and dissolving ticipated in a pilot test that helped
process and then select the maps” (Golledge 2002). estimate the difficulty level of
appropriate map layers involved STAT and reduce the incidence
in the overlay. (See Fig. 1)
of errors in test administration.
(Continued on next page)
19
 Jongwon Lee and Robert Bednarz

Revisions were made as a conse-

Table 2. Description of question types and spatial thinking components to measure. quence of the pilot tests: questions
(Continued) that were perceived to have more
than one correct answer or were
difficult to score objectively were
Type Item Spatial Thinking
eliminated or revised; a pair of
(Item Number) Description Components to Measure
questions (for pre- and post-test)
VIII (#13, #14, In order to solve item #13, #14, Item #13, #14, #15, and #16 that proved to have different levels
#15, #16) #15, and #16, participants measure the trait of difficulty were adjusted; and
should visually extract types of “comprehending integration of items whose graphics or directions
spatial data from verbally geographic features were unclear were improved.
expressed spatial information. represented as points, When workshops to review
(See Fig. 1) networks, and regions” the TGMG project materials were
(Golledge 2002) and
scheduled during a variety of ge-
“comprehending spatial shapes
ography meetings and conferences,
and patterns” (Golledge 2002).
STAT review sessions were also
conducted. For example, a STAT
review session occurred in con-
junction with the annual meeting
of the National Council for Geo-
Downloaded by [66.31.30.1] at 19:41 02 January 2015

graphic Education. During that ses-

sion twenty-two TGMG workshop
participants including a preservice
teacher, education students, sec-
ondary school geography teachers,
and professional geographers, took
and commented on the STAT.

TEST RESULTS

Reliability and Construct Validity

of STAT
Test results from 352 university
students from four different U.S.
states who took STAT were used to
examine the reliability and validity
of STAT. The number of students
who completed the tests at the four
universities varied from 11 to 146.
The variation of the sample sizes
resulted from the access to students
by faculty volunteers who agreed
to administer STAT at each of the
schools. As a measure of inter-
nal consistency, Cronbach’s alpha
was calculated. Cronbach’s alpha
is a measure of the intercorrelation
of items, measuring the extent to
which item responses obtained at
the same time correlate with each
other. A value of 0.7 for Cronbach’s
alpha is generally considered to
Figure 1a. Selected items from the STAT. Each item corresponds to Type I, V, and indicate a reliable set of items (de
VI, respectively. Vaus 2002). The Cronbach’s alpha
for the latest version was 0.721
and 0.701 for Forms A and B,

20
 Components of Spatial Thinking: Evidence from a Spatial Thinking Ability Test

In order to explore to what ex-

tent spatial thinking skills are com-
posed of distinct components, the
construct validity of STAT was ex-
amined using factor analysis. Fac-
tor analysis is a statistical tech-
nique used to identify the minimal
underlying factors needed to ex-
plain the intercorrelations among
the test items. Principal compo-
nents analysis revealed six factors
with eigen-values of 1.0 or more,
accounting for 54.66 percent of
cumulative variance. In general, a
factor analysis accounting for 60–
70 percent or more of the total
variance is considered a good fit
to the data. Varimax rotation was
then applied to the six factors.
This procedure rotates the set of
Downloaded by [66.31.30.1] at 19:41 02 January 2015

individual scores within the space

defined by principal component
axes, thereby creating a new set
of factor loadings (increasing the
difference between high and low
loadings) for the factors that have
already been found. The rotated
factor matrix is presented in Table
3. The nature of each of the factors
in Table 3 is determined by the
characteristics of the variables that
have high loadings on these factors.
Six factors accounted for 11.1, 10.7,
10.5, 7.9, 7.6, and 6.9 percent of the
variance, respectively.
Figure 1b. Selected items from the STAT. Each item corresponds to Type If the skills tested by the eight
VII and VIII, respectively. Note: all items of STAT may be viewed at question types displayed in Table 2
http://home.ewha.ac.kr/∼ziriboy/STAT.pdf. are independent components of
spatial thinking, we would expect
the factor analysis to yield factors
respectively. When STAT was administered for the first that reflect those components. That is, questions that
time, we were somewhat disappointed by the relatively assess a specific component should be grouped. Although
low validity and reliability statistics. As we reconsidered the some factors directly or indirectly show high levels of
results, however, we realized that recent conceptualizations correspondence, others do not. For example, four of the
of spatial thinking skills support the notion that spatial questions that load on factor 2, items #13, #14, #15, and
thinking skills are composed of several elements that #16, are based on the spatial skill “visually extract types of
may be at least somewhat independent of one another. spatial data from verbally expressed spatial information”
Therefore, it is not surprising that some individuals might (Type VIII) although #10, which is not a Type VIII item,
perform significantly differently on questions that assessed also loads as highly on factor 2 (0.484). Factor 3 generates
different skills thereby lessening the internal consistency or heavy loadings for the items related to question type VII
intercorrelation. Although originally Golledge and Stimson (#9, #10, #11, and #12) requiring “participants to visually
(1997) proposed “spatial relations” as an additional spatial verify a map overlay process and then select the appropriate
ability to visualization or orientation, it seems likely that map layers involved in the overlay.” It is interesting that,
“spatial relations” included a variety of skills that are likely in addition to the type VII questions, item #7 had the high
uncorrelated. loading (0.529) on this factor. This result might be explained

21
 Jongwon Lee and Robert Bednarz

load equally on more than one

Table 3. Results of the factor analysis of the pretest of the STAT. factor such as item #8, which loads
on four factors—1 (0.292), 2 (0.207),
4 (0.195) and 5 (0.288)—although
Question at relatively low levels. This may
Item Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6
occur partly because the question
#4 .742 .126 .067 .236 −.110 .063 items are inadequately specified
#5 .682 .134 .110 .118 .052 −.237 (i.e., they represent or require more
#12 .560 −.060 .189 −.190 .084 .235 than one spatial thinking skill) or
#14 .407 .399 .074 .020 .130 .271 because of the failure of STAT to
#15 .089 .773 .002 .042 −.016 .156 capture the full range of spatial
#13 .054 .615 .041 .351 −.189 −.153 thinking.
#10 −.032 .484 .470 .087 .343 −.072
#16 .250 .451 .383 −.275 .195 .048
#9 .105 .113 .694 −.022 .157 .164 Analysis of STAT Test Results
#11 .063 .065 .674 .262 −.290 .071 The STAT was administered to
#7 .299 −.094 .529 .043 .024 −.280 students at a wide range of aca-
#3 .119 .118 .049 .687 .042 −.023 demic levels—at four universities
#2 .040 .027 .166 .591 .243 .388 located in Texas, Ohio, Illinois, and
#1 .024 −.039 .020 .101 .857 .039 Oregon, and at a junior high and
#6 .066 .103 .033 .104 .015 .695 high school in Ohio (Table 4).
.292 .207 .044 .195 .288 −.332
Downloaded by [66.31.30.1] at 19:41 02 January 2015

#8 In general, as students advanced

Eigen values 1.781 1.705 1.676 1.265 1.218 1.101 from junior high to university their
% of variance 11.129 10.659 10.472 7.908 7.613 6.879
performance improved. For every
Cumulative % 11.129 21.788 32.260 40.168 47.781 54.660
question, the average score for high
school students exceeded that for
by the existence of similar cognitive processes in solving junior high school students. Similarly, the average scores
type VII (“visually verify map overlay processes”) and item for university students were greater than the scores for high
#7 (“comprehending spatial association”). school students for every question, although the average
However, unlike factors 2 and 3, other factors were high school scores were greater for some questions than the
not clearly connected to specific question types. Factor average scores of university B and C students.
1, with heavy loadings for items #4, #5, #12, and #14, Analysis of the test data showed that a sizeable majority
spans several question types. Furthermore, some items of university students could identify patterns on a map
and choose a correct graphical
display of a spatial pattern (item
#3). In addition, a large percent
Table 4. A percentage of correct answers per item and mean score by groups.
of university students (63.6 to 93.1
percent) could find locations, un-
Univ. A Univ. B Univ. C Univ. D Junior High High derstand orientations and direc-
Item (N = 29) (N = 11) (N = 59) (N = 146) (N = 52) (N = 149) tions, and navigate on road maps
following directions (item #1 and
#1 89.66 90.91 67.80 84.25 42.27 73.51
#2). Whereas nearly 90 percent of
#2 93.10 63.64 77.97 88.36 43.75 67.58
the participants could comprehend
#3 96.55 81.82 91.53 95.21 46.32 79.78
#4 65.52 45.45 54.24 70.55 24.21 38.46 spatial association between two
#5 65.52 45.45 54.24 79.45 19.35 29.89 maps (item #6), less than half were
#6 96.55 100.00 83.05 94.52 55.32 71.27 capable of transforming a spatial
#7 65.52 27.27 18.64 42.47 22.34 24.73 relationship into a graphic form
#8 51.72 27.27 28.81 44.52 18.48 24.73 (item #7). Although items #9, #10,
#9 68.97 36.36 52.54 60.27 26.37 30.34 #11, and #12 were designed to
#10 79.31 54.55 67.80 82.19 31.91 60.44 assess the same skill (Type VII),
#11 65.52 27.27 49.15 58.90 15.22 33.71 students performed most poorly on
#12 34.48 18.18 22.03 32.19 15.22 24.72 item #12 (see Fig. 1b), the question
#13 93.10 45.45 64.40 75.34 35.16 55.25
that required the most complex
#14 68.97 36.36 55.93 67.12 30.53 37.36
Boolean logic.
#15 89.66 54.55 72.88 86.99 41.94 63.13
#16 68.97 9.09 45.76 69.18 22.83 36.11 A few question items were
Mean Score 11.93 7.64 9.07 11.32 4.60 7.58 found to have better discriminat-
(SD) (2.64) (3.67) (2.79) (2.82) (2.47) (2.76) ing power than other items. For
instance, question items such as #1,

22
 Components of Spatial Thinking: Evidence from a Spatial Thinking Ability Test

Analysis of variance (ANOVA)

was used to determine whether
significant differences in scores
among four university groups ex-
isted. When analysis of variance
(ANOVA) was performed on Form
A scores of the four schools, signif-
icant differences were found ( p =
.000) (Table 5). Post-hoc compar-
isons (using Tukey method) re-
vealed that students of university
A and D scored significantly higher
than those of university B and C
(Table 6). No significant difference
was found between scores of uni-
versity A and D and between those
of university B and C, respectively.
The number of geography ma-
jors in each group may account for
Figure 2. Score comparison per item by groups. the differences in scores (Table 7).
Downloaded by [66.31.30.1] at 19:41 02 January 2015

School groups with the highest

percentages of geography majors,
#2, #3, #10, and #11 most clearly separate the junior high university A (41.38% majors) and D (26.03%), scored higher
schools students from others (Fig. 2). Nevertheless, as the on Form A than university B (9.09%) and C (16.95%).
graph illustrates, students at all levels displayed similar The format of STAT, including maps and spatial terms,
performance patterns, in the sense that scores for all may be more challenging or less familiar to nongeography
students were uniformly higher for some questions than majors.
others. The lines representing the scores for the six student Because two versions of STAT were developed (Form
samples are relatively parallel, indicating that students, A and Form B), it was important to verify that the
from junior high to university, scored higher and lower two forms were equally difficult. This was achieved by
on the same questions. This result would seem to indicate comparing participants’ mean scores on the two forms
that some skills are more challenging than others and offer of the STAT using a t-test for independent samples. The
support for the argument that spatial thinking is composed student participants of university A, B, and C (students of
of more than one skill or ability (in addition to the widely university D took only Form A) were randomly divided
accepted spatial visualization and orientation abilities). into two groups each taking either Form A or Form
B of the test. In order to verify the
equivalency of two forms of the
Table 5. ANOVA for Form A scores by groups (University students only). STAT, three separate t-tests were
conducted for the three groups
(Table 8). All of the t-tests indicated
SS df MSE F p that the two forms of STAT are
generally equivalent in difficulty
Between groups 359.524 3 119.841 14.936 .000
except for university B where two
Within groups 1933.643 241 8.023
Total 2293.167 244
students who completed Form A
scored very low on the test (an-
swering only two of sixteen ques-
tions correctly).
Table 6. Post-hoc comparison of Form A scores by groups.
DISCUSSION AND
CONCLUSIONS
Univ. A Univ. B Univ. C In this study, we developed
and evaluated standard measures
Univ. B 4.295** (.000)
of spatial thinking skills. Inter-
Univ. C 2.863** (.000) 1.431 (.416)
nal consistency reliability estimates
Univ. D 0.616 (.708) −3.679** (.000) −2.247** (.000)
for the STAT were in the moder-
** p < .01. ate range. Although these results
may raise concerns regarding the

23
 Jongwon Lee and Robert Bednarz

individuals engaged in spatial

Table 7. A comparison of Form A scores by major (geography major versus nongeography problem solving before this issue
major). can be addressed reliably.
Perhaps another reason that the
factor analysis did not identify
N Mean S.D. t p
eight independent components is
Geography major 61 11.77 2.58 −3.264 0.067 that independence of the eight
Nongeography major 184 10.32 3.14 components is not as great or as
complete as hypothesized. Spatial
thinking skills may be comprised
of fewer than eight components or
some skills may be correlated to
others, which may or may not be
Table 8. Independent sample t-test.
the same thing. If spatial think-
ing consists of fewer independent
Form N Mean S.D. t p components, what do the results
of the factor analysis suggest those
University A A 29 11.93 2.64 .908 .368 components might be? Three of the
B 26 11.23 3.08 four items loading highly on factor
University B A 11 7.64 3.67 −1.945 .065 1 require the skill to overlay or vi-
B 13 10.38 3.25
Downloaded by [66.31.30.1] at 19:41 02 January 2015

sualize spatial data. Four of the five

University C A 59 9.07 2.78 −0.771 .442
B 47 9.49 2.81
items that load on factor 2 require
the ability to distinguish among
the map elements point, line, and
completeness of the measure, there are several issues that area. Three of the items that load on factor 3 test respon-
may account for the moderate reliability. With regard to dents’ skill in performing Boolean operations on geometric
internal consistency reliability, Cronbach’s alpha increases pattern; the fourth item requires identification of the nature
as the number of items increases. Ceteris paribus, increasing of spatial correlation between two mapped distributions.
the number of items can increase the level of alpha. The two items loading on factor 4 do not appear to have
In terms of the construct validity, the factor analysis much in common: one concerns a way-finding task and the
using principal components extraction with varimax meth- other requires the creation of a cross-section diagram from a
ods provide mixed results with regard to the research mapped distribution. Factors 5 and 6 are, for the most part,
hypothesis. We hypothesized that factor analysis would comprised of one item, a way-finding question for factor 5
identify the independent components of spatial thinking and identification of a positive spatial correlation for factor
by generating factors that reflected the eight components of 6.
previous researchers’ spatial thinking conceptualizations Thus, the analysis of STAT offers relatively little support
that were represented by questions in STAT. Whereas some for the existence of the independent spatial thinking
factors were directly or indirectly connected to the question components hypothesized in the literature. The analysis
types, some were not. This result might be attributed to also suggests that Golledge and Stimson’s spatial rela-
the participants’ styles of spatial problem solving. It is tions ability is almost certainly not a single ability but
widely accepted that different people employ different instead is comprised of a collection of different skills.
strategies when solving spatial tasks (Kyllonen, Lohman, Based on the clusters identified by the factor analysis,
and Woltz 1984; Lohman and Kyllonen 1983). Furthermore, the following spatial thinking components emerge: map
spatial tasks are often solved using nonspatial processing visualization and overlay, identification and classification
strategies. For instance, Just and Carpenter (1985) found of map symbols (point, line, area), generalized or abstract
that many spatial test items may also involve verbal analytic Boolean operations, map navigation or way-finding, and
processing. They argued that verbal strategies are routinely recognition of positive spatial correlation. We do not assert
employed for spatial tasks including 3D rotation, spatial that these five components are the five spatial thinking skill
orientation, and others. Thus, for at least some individuals, sets. Nevertheless, intuitively these skills do seem different
relative success on spatial items could be due to a verbal enough that individuals might be able to use one or more
or another type of ability rather than this spatial ability. successfully while they are having difficulty with others.
As mentioned previously, care was taken when devel- For example, it is not hard to believe that a person who is
oping STAT to maximize spatial processes and minimize skilled at solving Boolean problems might not necessarily
verbal processes required to answer correctly. Because no be a skilled navigator.
information about how students solved questions was We do think that the analysis strongly supports the
collected however, studies with additional items are needed hypothesis that spatial thinking is a collection of different
to explore the spatial thinking processes employed by skills and that more work must be done to identify those

24
 Components of Spatial Thinking: Evidence from a Spatial Thinking Ability Test

component skill sets. The results also help explain why indi- Clements, D. H., M. T. Battista, J. Sarama, and S.
viduals perform well on some spatial thinking tasks while Swaminathan. 1997. Development of students’ spatial
performing poorly at others. For geography educators these thinking in a unit on geometric motions and area. The
results suggest that because students perform well on some Elementary School Journal 98 (2): 171–186.
tasks does not mean that they will perform well on others. If
Committee on Support for Thinking Spatially. 2006. Learn-
different tasks require different skill sets, performance may
ing to Think Spatially. Washington, D.C.: The National
be uneven. The results also suggest that giving students
Academies Press.
a variety of ways to demonstrate what they have learned
might reveal a student’s knowledge or ability that would go Dean, G. M., and P. E. Morris. 2003. The relationship
undetected if only one method of assessment is employed. between self-reports of imagery and spatial ability.
The two forms of STAT were equivalent in difficulty at British Journal of Psychology 94 (2): 245–273.
baseline, and therefore, the two forms of STAT can be used
DeMers, M. N., and J. S. Vincent. 2007. ArcAtlas in
for pre- and post-test designs to evaluate changes in spatial
the classroom: Pattern identification, description, and
thinking abilities over a brief period. In addition, the field
explanation. Journal of Geography 106 (6): 277–284.
tests in several different environments showed STAT was
useful for testing both university and high school students. de Vaus, D. 2002. Surveys in Social Research, 5th ed. London:
Our standardized measure needs to substantiate content Routledge.
validity more rigorously. This measure, however, provides
Doering, A., and G. Veletsianos. 2007. An integration of the
rigorous bedrock for testing that can be expanded with
use of real-time, authentic geospatial data in the K–1
new tests in the near future. The current version of STAT
classroom. Journal of Geography 106 (6): 217–225.
Downloaded by [66.31.30.1] at 19:41 02 January 2015

represents considerable developmental work based on

a solid theoretical foundation. Additional research and Eliot, J., and M. Y. Czarnolewski. 2007. Development of an
refinement of the measures could strengthen their testing everyday spatial behavioral questionnaire. The Journal
abilities and contribute to research on spatial literacy in the of General Psychology 134 (3):361–381.
long run.
Gersmehl, P. 2005. Teaching Geography. New York: The
Guilford Press.
REFERENCES Gersmehl, P. J., and C. A. Gersmehl. 2007. Spatial thinking
Albert, W. S., and R. G. Golledge. 1999. The use of spatial by young children: Neurologic evidence for early
cognitive abilities in geographic information systems: development and “educability”. Journal of Geography
The map overlay operation. Transactions in GIS 3 (1): 106 (5): 181–191.
7–21. ———. 2006. Wanted: A concise list of neurologically de-
Allen, T. R. 2007. Digital terrain visualization and virtual fensible and assessable spatial-thinking skills. Research
globes for teaching geomorphology. Journal of Geogra- in Geographic Education 8 (1) : 5–38.
phy 106 (6): 253–266. Gilmartin, P. P., and J. C. Patton. 1984. Comparing the
Association of American Geographers. n.d. Teachers’ Guide sexes on spatial abilities: Map-use skills, Annals of the
To Modern Geography. http://www.aag.org/cs/tgmg Association of American Geographers 74 (4): 605–619.
(last accessed November 2011). Golledge, R. G. 2002. The nature of geographic knowledge.
Battersby, S., R. G. Golledge, and M. J. Marsh. 2006. Annals of the Association of American Geographers 92 (1):
Incidental learning of geospatial concepts across grade 1–14.
level: Map overlay. Journal of Geography 105 (4): 139–146. ———. 2001. Geography and everyday life (again!). Di-
Bednarz, S. W. n.d. Maps and spatial thinking rections Magazine (March). http://www. directions-
skills in the AP human geography. AP Cen- mag.com/columns.php?column id=40 (accessed Au-
tral. http://apcentral.collegeboard.com/apc/public/ gust 2010).
courses/teachers corner/151317.html (accessed Au- ———. 2000. Geography and everyday life. Direc-
gust 2010). tions Magazine (December). http://www. directions-
Bednarz, S. W. 2004. Geographic information systems: A mag.com/columns.php?column id=37 (accessed Au-
tool to support geography and environmental educa- gust 2010).
tion? GeoJournal 60 (2): 191–199. ———. 1995. Primitives of spatial knowledge. In Cognitive
Carswell, R. J. B. 1971. The role of the user in the map com- Aspects of Human-Computer Interaction for Geographic
munication process: Children’s abilities in topographic Information Systems, ed. T. L. Nyerges, D. M. Mark,
map reading. Cartographica: The International Journal for R. Laurini, and M. J. Egenhofer, pp. 29–44. Boston,
Geographic Information and Geovisualization 8 (2): 40–45. Massachusetts: Kluwer Academic Publishers.

25
 Jongwon Lee and Robert Bednarz

———. 1992. Do people understand spatial concepts: The Kyllonen, P. C., D. F. Lohman, and D. J. Woltz. 1984.
case of first-order primitives. In Theories and Methods of Componential modeling of alternative strategies for
Spatio-temporal Reasoning in Geographic Space, ed. A. U. performing spatial tasks. Journal of Educational Psychol-
Frank, I. Campari, and U. Formentini, pp. 1–21. New ogy 76 (6): 1325–1345.
York: Springer-Verlag.
Lee, J., and R. S. Bednarz. 2009. The effect of GIS learning on
Golledge, R. G., M. Marsh., and S. Battersby. 2008. Match- spatial thinking. Journal of Geography in Higher Education
ing geospatial concepts with geographic educational 33 (2): 183–198.
needs. Geographical Research 46 (1): 85–98.
Lloyd, R., and R. L. Bunch. 2003. Technology and map-
Golledge, R. G., and R. J. Stimson. 1997. Spatial Behavior: A learning: Users, methods, and symbols. Annals of
Geographic Perspective. New York: The Guilford Press. the Association of American Geographers 93 (4): 828–
850.
Hall-Wallace, M. K., and C. M. McAuliffe. 2002. Design,
implementation, and evaluation of GIS-based learning Lohman, D. F., and P. C. Kyllonen. 1983. Individual
materials in an introductory geoscience. Journal of differences in solution strategy on spatial tasks. In
Geoscience Education 50 (1): 5–14. Individual differences in cognition, ed. D. F. Dillon and
R. R. Schmeck, pp. 105–135. New York: Academic
Harris, L. J. 1981. Sex variations in spatial skills. In Spatial
Press.
Representation and Behavior Across the Lifespan, ed. L. S.
Liben, A. H. Patterson, and N. Newcombe, pp. 31–54. Marsh, M., R. Golledge, and S. E. Battersby. 2007. Geospatial
New York: Academic Press. concept understanding and recognition in G6–college
Downloaded by [66.31.30.1] at 19:41 02 January 2015

students: A preliminary argument for minimal GIS.

Hegarty, M., A. E. Richardson, D. R. Montello, K. Lovelace,
Annals of the Association of American Geographers 97 (4):
and I. Subbiah. 2002. Development of a self-report
696–712.
measure of environmental spatial ability. Intelligence 30
(5): 425–447. Milson, A. J., and B. D. Earle. 2007. Internet-based GIS in an
inductive learning environment: A case study of ninth-
Janelle, D. G., and M. F. Goodchild. 2009. Location across
grade geography students. Journal of Geography 106 (6):
disciplines: Reflection on the CSISS experience. In
227–237.
Geospatial Technology and the Role of Location in Science,
ed. H. J. Scholten, N. van Manen, and R. v.d. Velde, Montello, D. R., L. L. Lovelace, R. G. Golledge, and C. M.
pp. 15–29. Dordrecht, Netherlands: Springer. Self. 1999. Sex-related differences and similarities in
geographic and environmental spatial abilities. Annals
Just, M. A., and P. A. Carpenter. 1985. Cognitive coordinate
of the Association of American Geographers 89 (3): 515–
systems: Accounts of mental rotation and individual
534.
differences in spatial ability. Psychological Review 92 (2):
137–172. Patterson, M. W., K. Reeve, and D. Page. 2003. Integrating
geographic information systems into the secondary
Kerski, J. J. 2008. The world at the student’s fingertips. In
curricula. Journal of Geography 102 (6): 275–281.
Digital Geography: Geo-spatial Technologies in the Social
Studies Classroom, ed. A. J. Milson and M. Alibrandi, Schultz, R. B., J. J. Kerski, and T. C. Patterson. 2008.
pp. 119–134. Charlotte, North Carolina: Information The use of virtual globes as a spatial teaching tool
Age. with suggestions for metadata standards. Journal of
Geography 107 (1): 27–34.
———. 2000. The implementation and effectiveness of
geographic information systems technology and meth- Wigglesworth, J. C. 2003. What is the best route? Route-
ods in secondary education. Ph.D. diss., University of finding strategies of middle school students using GIS.
Colorado, Boulder. Journal of Geography 102 (6): 282–291.

26