learning geographic information from a map and text: learning

Learning Geographic Information from a Map and Text:Learning Environment and Individual Differences

Robert Earl Lloyd and Rick L. BunchDepartment of Geography / University of North Carolina / Greensboro / NC / USA

Abstract

A map is frequently combined with a text to provide spatial and non-spatial information for learners. How a map anda text are combined and the characteristics of learners are keys for understanding successful learning. This study useda cognitive experiment to investigate spatial learning by explaining performance on a test of acquired knowledge withvariables related to the learning environment and to individual differences of learners. Results indicate that havingparticipants read a text beside a map produced the best performance. Participants were more successful at learningthe information in the text and less successful at learning the information on the map. Performance was measured byaccuracy, reaction time, and confidence measures; a standardized index for overall efficiency combined these measures.Performance was significantly related to individual difference variables measuring experience, verbal and spatial workingmemory capacity, 2D/4D digit ratio, and cognitive style. Sex and gender variables were not significantly related to varia-tions in performance. In complex learning situations, as in processing a combined map and text, the expected verbal andspatial processing advantages of female and male learners may both produce positive results. In more complex cases,variables related to brain asymmetry, memory capacity, and cognitive style may provide more useful explanations ofperformance.

Keywords: spatial learning, map, text, individual differences, memory capacity, digit ratio, cognitive style

Resume

On associe souvent du texte a une carte pour fournir des renseignements spatiaux et non spatiaux aux apprenants. Lamaniere dont la carte et le texte sont associes et les caracteristiques des apprenants sont des elements cles pour expliquercet apprentissage. Dans l’etude, on se sert d’une experience cognitive pour evaluer un apprentissage spatial et expliquerle rendement obtenu a un test sur les connaissances acquises en tenant compte des variables relatives au milieud’apprentissage et aux differences individuelles d’un apprenant a l’autre. D’apres les resultats, les meilleurs rendementssont obtenus quand les participants lisent le texte a cote de la carte. Les participants apprenaient mieux les renseigne-ments contenus dans le texte et moins ceux de la carte. Le rendement a ete mesure en fonction de la precision, du tempsde reaction et des mesures de confiance; un indice normalise de l’efficacite globale combinait ces mesures. Le rende-ment etait etroitement lie aux variables de la difference individuelle qui mesurait l’experience, la capacite de la memoireoperationnelle verbale et spatiale, le rapport numerique 2D/4D et le style cognitif. Les variables du sexe et du genren’etaient pas liees de pres aux variations du rendement. Dans des situations d’apprentissage complexe, comme lorsdu traitement d’une carte avec du texte, les avantages prevus relativement au traitement verbal et spatial pourraientproduire des resultats positifs tant pour les apprenants que pour les apprenantes. Dans des cas plus complexes, lesvariables liees a l’asymetrie du cerveau, a la capacite de la memoire et au style cognitif pourraient fournir une meilleureexplication du rendement obtenu.

Mots cles : apprentissage spatial, carte, texte, differences individuelles, capacite de la memoire, rapport numerique, style cognitif

Cartographica (volume 45, issue 3), pp. 169–184 doi: 10.3138/carto.45.3.169 169

Introduction

The interaction between a map and text has been studied

using different assumptions about the dominant source of

information. Some studies have focused on reading com-

prehension and whether the map can aid learning with

text (Kulhavy, Stock, and Kealy 1993). Michael Verdi and

Raymond Kulhavy (2002) argue that when sequentially

learned, the two are partners that work together to aid

the learning process. Others have suggested that the

combination can impede learning when they compete for

attention (Bunch and Lloyd 2006). Research on map–text

combinations needs to place basic questions on learning

performance within contexts that consider the effects of

the learning environment and the characteristics of map

learners.

Maps express spatial information in a variety of ways. A

reference map, for example, uses map symbols to illus-

trate spatial relationships and meaning. Narrative texts

provide additional descriptions of map features and their

relationships. This study uses a hypothetical map of an

island and three experimental methods for presenting

narrative text. The purpose is to identify which method

produces the most efficient performance when a novel

map and text are simultaneously available. The methods

used to present geographic information represent experi-

ences one might encounter when learning about an un-

familiar place in a book or on a Web page. In this study,

an efficient performance is defined as task scores indicating

high accuracy, high confidence, and fast response time.

Task scores that indicate low accuracy, slow response

time, and low confidence reflect a negative performance.

Memory for Objects and Locations

Cognitive scientists make a distinction between object

memory and location memory (Bellgowan and others

2009). Verbal processes in the left hemisphere encode

categorical spatial information, and visual processes in the

right hemisphere encode coordinate spatial information

(Kosslyn and others 1998). Categorical spatial informa-

tion is easier to learn but less exact; coordinate spatial

information is more accurate but also more difficult to

learn. Coordinate spatial information for a city indicates

its horizontal and vertical locations on a map, while cate-

gorical spatial information for the same city would relate

to its location in a specific region or along a particular

river. Both types of information indicate where the city is

on a map. Functional magnetic resonance imaging (fMRI)

can identify regions of the brain associated with cate-

gorical and coordinate visual memory (Slotnick and Moo

2006; van der Lubbe and others 2006). Although there are

separate systems for processing spatial and non-spatial

information, these types of information are associated

with one another in memory (Sommer, Schoell, and

Buchel 2008).

Although the typical learner is capable of encoding verbal

information using left-hemisphere processes and visual–

spatial information using right-hemisphere processes,

individuals may have these resources asymmetrically dis-

tributed within regions of the brain (Fenner, Heathcote,

and Jerrams-Smith 2000). Evidence suggests that a com-

plex mix of biological and environmental factors affects

an individual’s spatial abilities (Casey 1996). Individual

differences such as brain structure, gender, memory capaci-

ties, and cognitive styles are likely to affect performance in

learning and recalling geographic information. Individual

learners who favour verbal processing may be more effi-

cient when reading text, while individual learners who

favour visual-spatial processing may be more efficient

when reading a map.

Asymmetrical Spatial Learning

Asymmetrical brains provide an advantage for map readers

by representing characteristics of mapped objects and rela-

tionships among these objects simultaneously as category

and coordinate information (Jager and Postma 2003).

David Patton and Robert Lloyd (2009) found evidence

supporting asymmetrical differences between participants

who searched a small-scale map for well-known cities

and those who searched for novel cities. Their results

suggest that participants who searched for known cities

encoded coordinate information, while those who searched

for the novel cities encoded categorical information, and

demonstrated a sex-related bias in encoding tendencies

between female participants (categorical) and male parti-

cipants (coordinate). The current study further examines

asymmetrical learning using more complex map and text

combinations that present opportunities to learn both

spatial and non-spatial information.

Individual Differences of Map Learnersand Performance

Any study involving map use needs to take into con-

sideration the differences among the individuals who use

the map. In this section, we review evidence that gender,

working memory capacity, hand-digit ratios, and cogni-

tive styles are likely to influence spatial learning.

gender identity or sex?

Numerous cognition studies have used sex as a measure

of individual differences among participants (Halpern

2000; Kimura 2000). Some researchers have argued that

a classification based on sex is limiting and that gender

identification provides a more meaningful perspective for

cognition studies (Hardwick and others 2000; Lloyd and

Bunch 2008). Studies considering gender have demon-

strated that female participants who score high on the

Robert Earl Lloyd and Rick L. Bunch

170 cartographica (volume 45, issue 3)

masculinity scale also score higher on a variety of spatial

performance tasks (Signorella and Jamison 1986). It is also

true that extended practice on spatial tasks can enhance

overall performance and that female gender identity can

guide interests that lead to such practice (Lloyd, Hodgson,

and Stokes 2002; Casey 1996; Patton and Lloyd 2009).

Deborah Saucier and others (2002) found that gender-

role socialization mediates sex differences in spatial tasks

requiring mental rotation. Lloyd and Bunch (2008) used

the Bem Sex Role Inventory (Bem 1974) to measure

masculinity and femininity scales for participants viewing

a map of the US states; that study reported better per-

formance for participants who scored high on both the

masculinity and femininity scales, and higher accuracy

for those who scored lower on both the masculinity and

femininity scales. The current research considers the

influence of gender identity on spatial learning involving

map and text combinations.

working memory capacity

Alan Baddeley’s (2003) multi-component model of work-

ing memory contains a component for processing visual

information (the ‘‘visuospatial sketchpad’’) and a separate

component for processing verbal information (the ‘‘pho-

nological loop’’) that have a limited storage capacity.

Cognitive load theory (CLT) explains how people use

both visual-spatial and verbal information during learning

and problem solving (Sweller 1988; Bunch and Lloyd

2006). A key proposition of CLT is that working memory

has a limited capacity but connects to unlimited long-

term memory (Sweller 1988; Baddeley 1998). Tests of

working memory span partially explain variations in

performance on real-world cognitive tasks (Fischer 2001;

Cowan and Morey 2006). Lloyd and Bunch (2008) report

a significant relationship between performance accuracy

on a search task and the results of verbal- and spatial-

memory-span tests. The current study uses working-

memory-span tests to explore spatial learning with a map

and text.

hand-digit ratios and asymmetrical brains

Cognitive abilities affected by sex hormones have both

lifelong organizational and short-term activational effects

on behaviour (Kimura 1989; Falter, Arroyo, and Davis

2006). Scott Bell and Deborah Saucier (2004), based on

their study of activational effects, suggest that optimal

performance on a spatial task occurs with moderate levels

of testosterone; this moderate level is relatively low for

men and relatively high for women. Catherine Gouchie

and Doreen Kimura (1991) report the same pattern, with

higher-testosterone women and lower-testosterone men

exhibiting better spatial abilities.

Exposure to prenatal sex hormones affects asymmetries in

the human body, including how the brain is structured

(Manning and others 1998; Janowsky 2006). Researchers

consider the ratio of the second digit to the fourth digit

on the hand (the 2D/4D ratio) as a biomarker to identify

brain asymmetry. Prenatal testosterone slows the growth

rate of the left side of the brain while enhancing growth of

the right side (Brosnan 2006); the length of the fourth digit

(ring finger) is an index of prenatal testosterone exposure,

and the length of the second digit (index finger) is an index

of prenatal oestrogen exposure. The 2D/4D ratio, therefore,

provides an index of asymmetrical development of the

brain hemispheres. Men generally have a significantly

lower mean 2D/4D ratio than women (Loehlin and others

2006). Studies have also linked the 2D/4D ratio to the

variation of spatial abilities (Coolican and Peters 2003;

Csatho and others 2003; Falter and others 2006). Lloyd

and Bunch (2008) report results indicating that the 2D/4D

ratio interacted with both gender and working memory

capacity to explain reaction time and accuracy in a search

for target states on a US map.

cognitive styles—empathizing and systemizing

Another promising way of looking at individual differ-

ences has been termed ‘‘cognitive style.’’1 In broader

terms, cognitive style might be thought of as the way in

which a person approaches the world and solves prob-

lems. The origin of the ‘‘empathizing–systemizing’’ theory

of psychological sex differences is in autism-disorder studies

presented by Simon Baron-Cohen and colleagues. Em-

pathizing and systemizing are two different dimensions

measured by an empathy quotient (EQ) and a systemiz-

ing quotient (SQ) respectively (Baron-Cohen and others

2003; Baron-Cohen and Wheelwright 2004). People who

score high on the EQ have a cognitive style that drives

them to identify another person’s emotions and thoughts,

while people who score high on the SQ have a cognitive

style that drives them to analyse or construct systematic

relationships in non-social domains. Both women and

men show a range of variation on both EQ and SQ dimen-

sions, but women score on average higher than men in

empathizing and men higher than women in systemizing

(Nettle 2007). Other studies (Billington, Baron-Cohen,

and Wheelwright 2007; Focquaert and others 2007) have

reported that, on average, those interested in the humani-

ties favour an empathizing style while those interested in

sciences favour a systemizing style, regardless of sex.

Studies on cognitive style that connect spatial learning

with maps are rare. One exception (Billington, Baron-

Cohen, and Bor 2008) considered participants with a

range of scores on the SQ using an experimental task

that elicited a conflict by having participants search for a

target found at a global or local level (i.e., large letters

made from small letters). They expected that an indi-

vidual with a high systemizing style would be proficient

in analysing the rules of a system. They report that par-

ticipants with higher SQ scores appeared to focus their

Learning Geographic Information from a Map and Text: Learning Environment and Individual Differences

Cartographica (volume 45, issue 3) 171

attention on local details and performed the search task

with more success when the target was at the local level.

Map learners with high SQ scores may have more interest

in and experience with information displayed on maps

than map learners with high EQ scores do. The current

research considers the potentially interesting relationships

between cognitive styles and spatial learning using map

and text combinations.

Research Design

The cognitive experiment used in the study had an initial

learning phase followed by a testing phase (see Figure 1).

All participants considered the same information in the

text and map, but they used different learning methods.

Three learning methods presented the information in dif-

ferent ways through auditory and visual sensory channels.

The step after the learning phase tested participants on in-

formation available in the text and on the map. Accuracy,

Reaction Time, and Confidence measured the Efficiency

of performance.2 Accuracy is a traditional measure of

learning success, but Reaction Time and Confidence have

also been useful in understanding overall performance on

spatial tasks (Lloyd and others 2002; Lloyd and Bunch

2003).

The Method used by each participant and the partici-

pant’s Sex were recorded. Participants also responded

to tests that determined categories of Gender, working

Memory, and cognitive Style. The lengths of the index

finger (second digit) and ring finger (fourth digit) on the

right and left hands determined each participant’s 2D/4D

ratio.

participants

Ninety university students were paid $10 each to partici-

pate in the study. The participant pool was divided evenly

among female (n ¼ 45) and male (n ¼ 45) students. All

participants were from a university’s student population

and were 18 years of age or older.

level of experience

Participants were either placed in a lower experience cate-

gory by limiting their exposure to information about the

map and text or placed in a higher experience category

by increasing their exposure to information about the

map and text. A strategy called ‘‘test-enhanced learn-

ing’’ provided more experience (Roediger and Karpicke

2006). As Mark McDaniel, Henry Roediger, and Kathleen

McDermott (2007, 200) explain, ‘‘test-enhanced learning

refers to the fact that taking an initial test on studied

material enhances its later retention relative to simply

studying the material and then taking a final test.’’ Shana

Carpenter and others (2008, 438) have reported results

for learning linguistic knowledge that support the hypo-

thesis that ‘‘memory tests enhance learning and reduce

forgetting more than additional study opportunities do.’’

The method used in the current experiment was to have

an initial study session for both the lower- and higher-

experience groups; the higher-experience group also took

a practice test and then had a second study session. We

assumed that the practice test would provide a context

for the second study session that would enhance learning.

The statements on the practice test were different from

those on the final test but had a similar style and con-

tent. We assigned the two sets of statements randomly

to the practice and final tests to balance their effect on

performance.

the experimental map and related text

The experimental map showed a fictional isolated island

called the Isle of Dogs, bounded on the west by the Atlan-

tic Ocean and on the east by the North Sea (see Figure 2).

Figure 1. The basic structure of the learning experiment. Solid arrows represent the flow of information into and out ofparticipants’ brains; dashed arrows represent the flow of information within participants’ brains.



Other areas on the island were four named provinces,

shown in different hues. Each province was subdivided

into a number of unnamed counties, with named larger

cities and unnamed smaller cities. Other point symbols

for several cities represented economic activities. Linear

features on the map represented an unnamed highway

system and named rivers.

The narrative text with the accompanying map discussed

physical, cultural, and travel information associated with

the Isle of Dogs (see Appendix). The text supplied both

spatial and non-spatial information that in some cases

was also available on the map and in other cases was

novel information.

learning methods

Three learning methods exposed the participants to both

the map and the text, but in different ways. Study sessions

had no time limits and proceeded at a pace set by the par-

ticipant. Participants using the reading method had the

map (Figure 2) available on the left of the screen and the

text information on the right. Each study session pre-

sented the physical, cultural, and travel information in a

random sequence controlled by the participant. Partici-

pants using the reading method were constrained to read-

ing the physical, cultural, and travel information in only

two random sequences representing one study session.

Participants would use the visual channel to process infor-

mation from the map and the text (see Figure 1). CLT

argues that learning may be constrained if two sources of

information compete for a single channel under time con-

straints (Bunch and Lloyd 2006; Harrower 2007). Since

participants using the reading method were not under a

time constraint, they could visually and consciously attend

to either the map or the text.

Participants using the listening method had the map

(Figure 2) available on the left of the screen without any

visible text information on the screen. Instead, the phy-

sical, cultural, and travel text played through audio in a

random sequence controlled by selecting a ‘‘next’’ button.

Since each individual recording continued playing when

started by a participant, the information was available in

fixed blocks of time. As in the reading method, partici-

pants using the listening method could hear the recording

in only two random sequences per study session. The

Figure 2. Experimental map learned by participants using all learning methods



listening method represented a less traditional method,

but one that made the spatial and non-spatial information

available in two separate channels (see Figure 1). CLT

argues learning can be enhanced if information coming

from two sources can be provided in separate audio and

visual channels (Mayer and Moreno 2003; Bunch and

Lloyd 2006; Harrower 2007). The ability to focus visual

attention on the map and audio attention on the recorded

text simultaneously could potentially enhance learning.

Participants using the rollover method initially viewed

a map without text (Figure 2). The map, however, con-

tained rollover buttons that revealed text information

describing the physical, cultural, and travel characteristics

of particular locations. At the same time, a recording of

the displayed text played through audio. Participants could

control access by rolling over buttons in any sequence,

but they were limited to two sequences per button. The

rollover method created an interactive map whereby

participants could connect map locations with text infor-

mation provided separately through visual and audio

channels.

It has been reported that ‘‘difficulty in learning new con-

cepts can be alleviated by making text more cohesive which

makes readers less dependent on pre-existing knowledge’’

(Ozuru, Dempsey, and McNamara 2009, 239). Some re-

searchers have also suggested that non-linear text presen-

tations increase the cognitive load for audiences without

prior knowledge (Amadieu and others 2009; Calisir and

Gurel 2009). Since participants in this study had no prior

knowledge of the target content, the less cohesive rollover

method may have increased the cognitive load.

testing for knowledge learned

Participants evaluated statements presented on both the

practice and final tests as true or false. The knowledge

required to evaluate the statement was available with

equal frequency from the text only, from the map only,

or from both text and map. Some statements featured

non-spatial information, and some featured spatial infor-

mation. For example, the statement ‘‘The Isle of Dogs

was uninhabited before 1022 AD’’ features non-spatial

information that was only available from the text; the

statement ‘‘The city of Kenly is west of the Lundy River’’

features spatial information that was available only from

the map; and the statement ‘‘The Isle of Dogs has four

provinces’’ is based on information both on the map and

in the text. The initial statements had important featured

information missing. For example, ‘‘The Atlantic Ocean

is of the Isle of Dogs’’ would initially appear; the

participant clicked on the statement after reading it, and

the featured information ‘‘east’’ would then complete the

statement, allowing the participant to determine the true/

false evaluation. The reaction time clock started when

the featured information appeared and stopped when the

evaluation was completed. This procedure did not count

the time required to read the initial statement as part of

the response time.

Variables

performance variables

Accuracy, Reaction Times, and Confidence collectively

measured performance on the final test (see Table 1).

Figure 3. Percentage of participants in Female and Malecategories for the variables 2D/4D (a), Gender (b), Memory(c), and Style (d)



These variables previously have measured different dimen-

sions of performance with success (Lloyd and others 2002;

Lloyd and Bunch 2003, 2005; Bunch and Lloyd 2006).

We expected a positive performance to be accurate, fast,

and confident. We expected a negative performance to be

inaccurate, slow, and unconfident.

The experiment measured Reaction Time in milliseconds,

Accuracy as percent correct, and Confidence as ratings on

a scale of �100 to 100. These three variables combined

to compute a standardized measure of learning Efficiency

(E z). Fred Paas and Jeroen van Merrienboer (1993) provide

a method for computing and visualizing the relationship

between two measurements – mental effort and perfor-

mance. Their method computes z-scores for performance

and mental effort, which are used to derive an instruc-

tional efficiency score (E). E is the perpendicular distance

between the x, y coordinate locations of the z-scores and a

diagonal line representing where E is equal to 0.

We have adapted this method to include the computation

of z-scores for the three measurements of Accuracy (Az),

Reaction Time (RTz), and Confidence (Cz):

Ez ¼Az � RTz þ Cz

ffiffiffi

3p (1)

In this study, Ez was measured as the perpendicular dis-

tance between the data point for an observation and the

neutral line where Ez ¼ 0 along the x, y, and z axes. Effi-

cient learning increases as Ez becomes more positive and

decreases as it becomes more negative. The Ez score can

be used as a final measure of overall learning success by

combining and analysing Accuracy, Reaction Time, and

Confidence (see Figure 1 and Table 1).

variables related to the learning environment

The research design imposed three learning environment

variable on the participants (see Table 1) and assigned

each participant to a Method category. The variable

Experience represents the level of experience acquired

by each participant, and the Source of the information

indicates the origin of information.

individual difference variables

We measured individual differences among the parti-

cipants in categorical form to serve as main effects in a

statistical model (see Table 1). We recorded the Sex of

participants and also measured each participant’s response

to questions for the Bem Sex Role Inventory (Bem 1974)

to measure Gender identity (see Table 1). Performance

also relates to working memory capacity (Baddeley 2003).

We used span tests to measure verbal and spatial work-

ing memory capacities and assigned each participant to a

working Memory capacity category. Previous research has

related asymmetrical brain structure and a preference for

using verbal or spatial processes in problem solving to

the ratio of the second and fourth digits of the hand

(Manning 2002). We measured the lengths of the second

and fourth digits on both hands and computed the

average ratio for each participant; participants were then

Figure 4. Performance variable means for learningenvironment variables: learning Method (a), level ofExperience (b), and Source of information (c).



assigned to Low, Middle, or High categories based on their

2D/4D ratios (see Table 1). Simon Baron-Cohen (2002,

2003) has suggested connections among cognitive styles

related to empathy and systemizing, sex differences, and

task performance. We measured both the Empathy Quo-

tient (EQ) and the Systemizing Quotient (SQ) for each

participant and used the EQ and SQ scores to determine

a cognitive Style category for each participant (see Table 1).

Analyses

The first set of analyses considered the relationship between

Sex and the other individual difference variables: 2D/4D

ratio, Gender, working Memory capacity, and cognitive

Style. A multivariate analysis of variance (MANOVA) using

three performance measurements as the dependent varia-

bles and the learning environment and individual differ-

ence variables as main effects was the final analysis.

sex and other individual differences

Figure 3 illustrates associations between categories of Sex

(Female and Male) and categories for individual difference

variables. The association between Sex and 2D/4D cate-

gories was significant (w2 ¼ 9.6, p ¼ 0.008), and the dis-

tribution of Female and Male participants within the three

2D/4D categories suggests a meaningful pattern: Female

participants have High 2D/4D ratios, while Male partici-

pants have Low 2D/4D ratios (see Figure 3a), a pattern

supported by the findings of previous research (Loehlin

and others 2006). The relationship between the 2D/4D ratio

and spatial abilities, however, appears to be non-linear

when considering Sex. It has been hypothesized that men

and women in the Middle 2D/4D category would have

better performance scores on spatial tasks (Sanders, Sjodin,

and Chastelaine 2002; Falter and others 2006).

As one might expect, there was a significant relation-

ship between the Sex and Gender categories (w2 ¼ 15.2,

p ¼ 0.002). Although Female individuals were most fre-

quently associated with the Feminine category and Male

participants with the Masculine category, both sexes

appeared in all four Gender categories (see Figure 3b).

These results indicate a significant overlap between the

Sex and Gender categories.

Male and Female participants were distributed in a simi-

lar fashion within the working Memory categories (see

Figure 3c). The data do not support a Female advantage

for verbal working memory or a Male advantage for

spatial working memory (w2 ¼ 1.5, p ¼ 0.670).

Sex and cognitive Style were significantly related (w2 ¼14.8, p ¼ 0.002). Female participants were more fre-

quently represented in the Empathist and None categories,

while Male participants were more frequently represented

in the Systemist and Both categories (see Figure 3d). The

Systemist category had fewer Female participants, and the

Empathist category had fewer Male participants.

manova results

The learning Method used by participants was found to

be multivariate significant in explaining test performance

Figure 5. Performance variable means for individualdifference variables: 2D/4D ratio (a), working Memorycapacity (b), and cognitive Style (c).



(see Table 2). Between-subjects tests indicated that Method

was significant in explaining Accuracy and Confidence

but not in explaining Reaction Time (see Table 3). The

expected pattern for a positive performance would have

positive z-scores for Accuracy and Confidence and nega-

tive z-scores for Reaction Time. None of the learning

methods indicated this pattern (see Figure 4a). The

Accuracy of performance, however, strongly favoured

the Read Method over the Listen or Rollover Methods.

The Read Method also had the only positive mean for

the Confidence variable. The Reaction Time means were

not significantly different, but participants using the Read

Method had the slowest mean time. The Efficiency index

clearly points to the Read Method as superior to the Listen

or Rollover Methods (see Table 2).

There could be several explanations for the dominance of

the Read Method. First, participants were likely to have

used the Read method more frequently in past experi-

ences with books and Web sites. Second, the less familiar

Listen and Rollover methods had lower Accuracy and

Confidence means, even though separate audio and visual

channels were equally available (see Figure 4a); although

CLT argues that two sensory channels are likely to pro-

duce better results than one channel, this would be true

only if the map and text competed for attention in the

visual channel. With no time constraint, participants

could switch their attention between the text and map

to further consider any map locations mentioned in the

text, and this would prevent competition in the visual

sensory channel. With a time constraint, switching atten-

tion would not be possible, and competition in the visual

sensory channel would increase the cognitive load. Third,

the Rollover method presented text information within a

non-linear structure that may have increased the cognitive

load by substantially reducing the coherence of the narra-

tive (Seufert 2003). The opportunity to look at the map,

read text, and listen to text simultaneously may also have

produced a general overload of information that reduced

comprehension.

Main effect Experience was multivariate significant (see

Table 2) and was univariate significant only for the

Confidence variable (see Table 3). A positive performance

pattern was found for the High Experience category, while

a weak negative performance pattern was found for the

Low Experience category (see Figure 4b). The z-scores

for Accuracy and Confidence were positive for the High

Table 1. Summary of the variables used in the analyses and expected relationships

Variable Measurement Form Expectation

Dependent Continuous Measures learning performance

Accuracy Percentage of correct responses Positively related to performance

Reaction Time Milliseconds Negatively related to performance

Confidence Scale (�100 to þ100) Positively related to performance

Efficiency (Ez) Standardized combined effects ofAccuracy, Reaction Time, and Difficulty

Positive values ¼ more accurate, faster, andmore confidence

Negative values ¼ less accurate, slower, andless confidence when recalled

Independent Categorical Expectation

Learning Method Listen, Read, Rollover Not clear which method is most effective

Source of Information Map, Text, Both Advantage for Both

Experience Low, High Performance improves with experience

Sex Female, Male Male advantage on tests for geographicknowledge; Female advantage on object-location memory

2D/4D Ratio Low, Middle, High Spatial ability advantage for Middle

Gender Androgynous, Masculine, Feminine,Undifferentiated

Masculine advantage on map Feminineadvantage on text

Memory Dual, Spatialist, Verbalist, Neither Working memory capacity positively relatedto performance

Cognitive Style Both, Systemist, Empathist, None Systemist advantage for maps



category and negative for the Low category, but Reaction

Time means were virtually zero for both Low and High

Experience categories. The Efficiency index indicated a

relatively small advantage of High over Low Experience

(see Table 2). It is no surprise to find that added Experi-

ence resulted in elevated levels of performance, especially

since other studies have also found that practice tests

enhance learning (McDaniel and others 2007).

Main effect Source was marginally multivariate significant

(see Table 2). Between-subjects tests indicated a univariate

significant difference only for the Accuracy variable (see

Table 3). A positive performance pattern was found for

the Text category, while the Map category showed a nega-

tive performance pattern (see Figure 4c). The Both cate-

gory means had a positive z-score for Confidence and a

negative z-score for Reaction Time, but the Accuracy

mean had the wrong sign for a positive performance. The

Efficiency index indicates Text as the best Source for

information and the Map as the worst Source (see Table 2).

Learners appear to have been able to glean information

from the Text more successfully than from the Map. The

salience of the information in the Text and on the Map

may explain this pattern. Participants may have con-

sidered specific information in the text important to learn

simply because it was stated explicitly. For example, the

text explicitly states that the first settlement on the island

was in AD 1022 (1022 CE). The relative importance of

specific information on the map may be less obvious. For

example, none of the rivers flow through Nussex; this

information is on the map, but not in the text, and a

participant may not even notice this fact or consider it

important.

Individual difference effects are 2D/4D ratio, working

Memory capacity, cognitive Style, Sex, and Gender. The

main effect 2D/4D ratio was multivariate significant (see

Table 2). Between-subjects tests indicated a univariate

significant difference only for the Confidence variable

(see Table 3). None of the 2D/4D ratio categories indi-

Table 2. Multivariate significance test for significant differences using standardized Accuracy, Reaction Time, andConfidence as simultaneous dependent variables

Effect Pillai’s TraceF Statistic

Probability>F

Efficiency*Ez

Learning Method 14.1 0.000 Listen ¼ �0.261Read ¼ 0.536Rollover ¼ �0.262

Level of Experience 3.3 0.020 Low ¼ �0.186High ¼ 0.199

Source of information 2.1 0.051 Map ¼ �0.265Text ¼ 0.180Both ¼ 0.085

Sex 0.9 0.455 Female ¼ �0.162Male ¼ 0.166

2D/4D Ratio 2.2 0.042 Low ¼ 0.031Middle ¼ �0.227High ¼ 0.219

Working Memory capacity 5.3 0.000 Dual ¼ 0.231Spatialist ¼ 0.058Verbalist ¼ �0.552Neither ¼ 0.117

Gender 1.8 0.064 Androgynous ¼ 0.043Masculine ¼ 0.093Feminine ¼ �0.199Undifferentiated ¼ 0.233

Cognitive Style 3.0 0.002 Both ¼ 0.137Systemist ¼ 0.276Empathist ¼ �0.217None ¼ �0.152

* The highest positive Efficiency value for each category relating to each effect is highlighted in boldface type.



cated a positive or negative performance pattern (see

Figure 5a). The best Efficiency score was associated with

the High 2D/4D category (see Table 2).

The significance of the 2D/4D ratio effect produced an

unexpected result (see Table 2): the best overall perfor-

mance on the final test was associated with the High

category. The 2D/4D ratio correlates with an asymmetri-

cal development of the processes in the right and left

hemispheres of the brain through exposure to prenatal

hormones. Earlier studies have suggested a non-linear

relationship between the 2D/4D ratio and spatial abilities,

leading to the expectation that the Middle category might

be associated with superior task performance. The current

experimental task involved learning information rather

than performing a perceptual task; the association of the

High 2D/4D ratio category with higher levels of perfor-

mance would appear to indicate that verbal processes in

the left hemisphere were more useful for the experimental

task.

The main effect for working Memory capacity was multi-

variate significant in explaining Accuracy, Reaction Time,

and Confidence (see Table 2). Between-subjects tests in-

dicated a univariate significant difference for the Reaction

Time and Confidence variables (see Table 3). None of the

working Memory capacity categories indicated an ideal

positive or negative performance pattern (see Figure 5b).

The Efficiency scores produced an interesting pattern for

the working Memory capacity categories (see Table 2):

the Verbalist category had the only negative value.

Working Memory capacity was independent of Sex and

had the strongest effect on performance (see Figure 3 and

Table 2). The Dual category was associated with the most

positive overall performance on the final test. Since parti-

cipants had better-than-median Spatial and Verbal work-

ing Memory capacity, this might be expected. The very

negative performance associated with the Verbalist cate-

gory connected to a very slow mean Reaction Time and

a very low mean Confidence (see Figure 4). It is possible

that a higher capacity for working with verbal information

and a lower capacity for working with spatial informa-

tion resulted in a tendency for these participants to focus

their attention on processing verbal information, which

may have been a disadvantage in a learning task involving

a map and large amounts of spatial information.

The main effect cognitive Style was also multivariate sig-

nificant (see Table 2). Between-subjects tests indicated a

univariate significant difference only for the Confidence

variable (see Table 3). The Systemist Style indicated the

pattern for a positive performance, and the Empathist

Style a pattern for a negative performance (see Figure

Table 3. Univariate significance between-participants test for significant differences using standardized Accuracy,Reaction Time, and Confidence individually as dependent variables.

Effect Accuracy FStatistic(Probability > F )

Reaction Time FStatistic(Probability > F )

Confidence FStatistic(Probability > F )

Learning Method(Listen, Read, Rollover )

49.0 (0.000) 0.327 (0.721) 4.7 (0.010)

Level of Experience(Low, High )

2.4 (0.125) 0.001 (0.978) 9.3 (0.003)

Source of information(Map, Text, Both )

3.8 (0.023) 2.3 (0.101) 1.2 (0.289)

Sex(Female, Male )

0.114 (0.736) 0.141 (0.708) 2.6 (0.109)

2D/4D Ratio(Low, Middle, High )

0.8 (0.464) 0.054 (0.947) 6.6 (0.002)

Working Memory capacity(Dual, Spatialist, Verbalist, Neither )

0.84 (0.472) 8.9 (0.000) 8.1 (0.000)

Gender(Androgynous, Masculine, Feminine,Undifferentiated )

1.2 (0.315) 2.3 (0.076) 1.6 (0.186)

Cognitive Style(Both, Systemist, Empathist, None )

1.5 (0.227) 1.2 (0.300) 5.6 (0.001)



5c). These two categories were also associated with the

most positive and negative Efficiency scores (see Table 2).

Cognitive Style was a significant effect, and, as expected,

the Spatialist category was associated with the best positive

performance (see Table 2). The means for the Spatialist

category also showed a pattern of high Accuracy, fast

Reaction Time, and high Confidence (see Figure 5c).

This result supports the argument that cognitive style

may play an important role in models explaining spatial

learning.

The remaining individual difference effects, Sex and

Gender, were not multivariate significant (see Table 2).

Between-subjects tests also indicated no univariate signifi-

cant differences for Sex or Gender categories (see Table 3).

The nature of the learning task is one possible explanation

for this lack of significance: several studies have shown a

male advantage for naming places on world maps and a

female advantage for tasks involving the recall of object

locations in a spatial array (Dabbs and others 1998; Voyer

and others 2007). The tasks for this experiment required

learning both spatial and non-spatial information, very

different from the previous studies, which have focused

on isolating cognitive processes by task and sex. Other

studies have also demonstrated a male advantage on visual

tasks and a female advantage on verbal tasks (Halpern

2000; Kimura 2000). In this study, participants needed to

use both visual and verbal processes to acquire informa-

tion from the text and map. The expected advantages for

female and male learners may ultimately have balanced

out in considerations of overall performance because

each group was required to consider both types of infor-

mation. Gender was marginally significant (see Table 2).

It is interesting to find that the Gender category Undiffer-

entiated showed the most positive Efficiency score and the

best overall performance.

Conclusions

Both learning-environment and individual-difference varia-

bles affected map learning. We assumed that the ideal

positive performance would be accurate, fast, and confi-

dent; for the eight ‘‘winning’’ categories having the highest

Efficiency for each effect (see Table 2), only half had col-

lective ideal z-score mean patterns (i.e., positive values for

Accuracy, negative values for Reaction Time, and positive

values for Confidence). Three of the four that did not

show the ideal pattern (Read Method, High 2D/4D ratio,

and Undifferentiated Gender) had relatively slow Reaction

Time means. Although it seems reasonable to consider

an accurate, fast, and confident performance as the ideal

positive pattern, accurate and confident performances

also occur with slower reaction times. This might be the

case if the information needed to evaluate a statement

as true or false is voluminous, not well organized in

memory, or not completely coded. Reaction times for

simple tasks such as searching visual displays or for com-

plex athletic performances are known to decrease with

practice (Wilson, MacLeod, and Muroi 2008; Yarrow,

Brown, and Krakauer 2009). The time needed to search

one’s memory for the information learned from the

experimental map and text is also likely to decrease with

more experience.

There was no clear expectation as to which learning

Method would produce the best performance on the

final test. The more traditional method of acquiring new

information (i.e., reading a text with a map) clearly pro-

duced the best performance (see Table 2 and Figure 4a).

A familiarity advantage over other methods may partially

explain the success of the traditional method, but its suc-

cess may also be related to constraint differences between

the Read and Listen Methods and coherence differences

between the Read and Rollover Methods.

Experience and test performance should be positively

related. This common-sense expectation was verified (see

Table 2 and Figure 4b). The significant difference between

the two Experience categories also supports the use of a

practice test to boost learning when study time is limited.

It was expected that the Both category would be the

Source of information associated with the best perfor-

mance. This expectation stemmed from the idea that

redundant information in the text and on the map would

increase the likelihood of successful learning. The Both

category did have a positive Efficiency value, so this

expectation may be correct; however, the Text Source

had the highest Efficiency value. A simple explanation for

this advantage relates to the nature of information in the

Text-only category. Information available in the Text and

on the Map was spatial information, while information

available only in the Text was non-spatial informa-

tion. Learning non-spatial information is apparently easier

than learning redundant spatial information. The diffi-

culty of learning spatial information may also explain the

low Accuracy, low Confidence, and slow Reaction Time

means for the Map Source category (see Figure 4c).

Finally, participants may have found it easier to judge

what was worth learning from the text than to decide

what was worth learning from the map.

In general, women have an advantage in processing verbal

information while men have advantage in processing

visual-spatial information. Men also tend to score higher

on tests of geographic knowledge, while women are better

at learning the locations of objects in a space (Zinser,

Palmer, and Miller 2004; Voyer and others 2007). Sex,

however, was not a multivariate significant effect explain-

ing performance in the present study (see Table 2).

Because the information to learn was both on a map

and in a text, the reported female and male advantages

on previous isolated tasks may have been neutralized.



The literature indicates that men with higher-than-

average 2D/4D ratios and women with lower-than-

average 2D/4D ratios perform better on some spatial

tasks, such as mental rotation and navigation (Coolican

and Peters 2003; Csatho and others 2003). An expectation

that participants in the Middle of the 2D/4D ratio distri-

bution would also perform better on the current experi-

mental learning task proved not to be correct. Individuals

with High 2D/4D ratio have asymmetrical brains favour-

ing systems in the left hemisphere. This provided a clear

advantage in learning the map and text. Systems pro-

cessing verbal and categorical spatial information were

more important for the current learning task than right-

hemisphere systems that support perceptual tasks. This

finding indicates that asymmetrical brains may have an

advantage for a range of spatial tasks, but not always in

the same way.

Gender was not a multivariate significant effect in con-

sidering performance variables (see Table 2). The same

argument as to why Sex was not significant may also

apply to Gender: if participants in the Masculine cate-

gory had better spatial abilities while participants in the

Feminine category had better verbal abilities, the separate

advantages may have cancelled each other out for the

complete task. The fact that the Undifferentiated Gender

category showed the highest Accuracy mean also supports

this notion.

Larger working Memory capacity should aid performance.

This hypothesis was generally supported for the current

task, as the Dual category had the highest Efficiency score

(see Table 2). Given that both spatial and non-spatial ver-

bal information were available to learn, it seems reason-

able that additional verbal and spatial working Memory

capacity would be beneficial.

The nature of the experimental task suggested that a cog-

nitive Style that focuses on understanding how systems

work would be superior to a cognitive Style that focuses

on other people’s emotions and thoughts. The multi-

variate significance of the cognitive Style effect and the

high Efficiency score for the Systemist category support

this expectation (see Table 2).

Experimental research that considers environments that

reflect real situations faced by map users should be

interesting to those interested in how people learn with

maps. Future studies should consider a range of learning

methods and how such methods interact with various

individual-difference variables and spatial tasks. For ex-

ample, tracking eye movements could supply data on

what type of information attracts attention and for how

long (Holsanova, Holmberg, and Holmqvist 2009). The

current study focused on examining explicit map-learning

tasks that closely resemble the learning environment

found in most typical classrooms. Future studies might

expand this contribution to include an examination of

how the individual differences of learners and map tasks

influence implicit learning.

Experimental research should also interest those con-

cerned with the design of maps for use by map readers

with special needs (e.g., tactile maps). It should also be

possible to design maps for map readers with special

advantages. Considering the effects of individual differ-

ences on performance will help in identifying less obvious

needs and advantages for particular map types. In the

future, it may be possible to systematically design maps

tailored specifically to meet the needs of targeted age

groups, map-reading levels, or cognitive styles. Maps

designed to fit the cognitive profiles of users could pro-

vide more certainty about their effectiveness. The cycle of

experimentation, exploration of instructional methods,

assessment of performance, and modification of map

designs would eventually lead to improved instruction,

map learning, and map-design protocols.

Author Information

Robert Earl Lloyd, Adjunct Professor, Department of

Geography, University of North Carolina, Greensboro,

NC 27412 USA. E-mail: [email protected].

Rick L. Bunch, Associate Professor, Department of Geo-

graphy, University of North Carolina, Greensboro, NC

27412 USA. E-mail: [email protected].

Notes

1. The term ‘‘style’’ is used in other contexts – for example,learning style based on a preference for using visual orverbal information (Graf, Lin, and, Kinshuk 2008).

2. Names of variables in the text appear in bold type;categories of these variables appear in italics. For exam-ple, Female and Male are categories of the variable Sex.

References

Amadieu, F., T. van Gog, F. Paas, A. Tricot, and C. Marine.2009. ‘‘Effects of Prior Knowledge and Concept-Map Struc-ture on Disorientation, Cognitive Load, and Learning.’’Learning and Instruction 19: 376–86.doi:10.1016/j.learninstruc.2009.02.005

Baddeley, A. 1998. Human Memory. Boston: Allyn & Bacon.

———. 2003. ‘‘Working Memory: Looking Back and LookingForward.’’ Nature Reviews Neuroscience 4: 829–39.doi:10.1038/nrn1201

Baron-Cohen, S. 2002. ‘‘The Extreme Male Brain Theoryof Autism.’’ Trends in Cognitive Sciences 6/6: 248–54.doi:10.1016/S1364-6613(02)01904-6

———. 2003. The Essential Difference: Men, Women and theExtreme Male Brain. London: Penguin.

Baron-Cohen, S., J. Richler, D. Bisarya, N. Gurunathan, andS. Wheelwright. 2003. ‘‘The Systemizing Quotient: AnInvestigation of Adults with Asperger Syndrome or High-Functioning Autism, and Normal Sex Differences.’’ Philo-sophical Transactions of the Royal Society of London,Series B: Biological Sciences 358: 361–74. doi:10.1098/rstb.2002.1206



Baron-Cohen, S., and S. Wheelwright. 2004. ‘‘The EmpathyQuotient: An Investigation of Adults with Asperger Syn-drome or High Functioning Autism, and Normal Sex Differ-ences.’’ Journal of Autism and Developmental Disorders 34:164–75. doi:10.1023/B:JADD.0000022607.19833.00

Bell, S., and D. Saucier. 2004. ‘‘Relationship among Envi-ronmental Pointing Accuracy, Mental Rotation, Sex, andHormones.’’ Environment and Behavior 36: 251–65.doi:10.1177/0013916503251470

Bellgowan, P., A. Buffalo, J. Bodurka, and A. Martin. 2009.‘‘Lateralized Spatial and Object Memory Encoding in Ento-rhinal and Perirhinal Cortices.’’ Learning and Memory 16:433–38. doi:10.1101/lm.1357309

Bem, S. 1974. ‘‘The Measurement of Psychological Androgyny.’’Journal of Consulting and Clinical Psychology B 42: 155–62. doi:10.1037/h0036215

Billington, J., S. Baron-Cohen, and D. Bor. 2008. ‘‘Systemiz-ing Influences Attentional Processes During the NavonTask: An fMRI Study.’’ Neuropsychologia 46: 511–20.doi:10.1016/j.neuropsychologia.2007.09.003

Billington, J., S. Baron-Cohen, and S. Wheelwright. 2007.‘‘Cognitive Style Predicts Entry into Physical Sciences andHumanities: Questionnaire and Performance Tests of Em-pathy and Systemizing.’’ Learning and Individual Differences17: 260–68. doi:10.1016/j.lindif.2007.02.004

Brosnan, M. 2006. ‘‘Digit Ratio and Faculty Membership:Implications for the Relationship between Prenatal Testos-terone and Academia.’’ British Journal of Psychology 97:455–66. doi:10.1348/000712605X85808

Bunch, R.L., and R.E. Lloyd. 2006. ‘‘The Cognitive Load ofGeographic Information.’’ Professional Geographer 58: 209–20. doi:10.1111/j.1467-9272.2006.00527.x

Calisir, F., and Z. Gurel. 2009. ‘‘Influence of Text Structureand Prior Knowledge of the Learner on Reading Compre-hension, Browsing and Perceived Control.’’ Computers inHuman Behavior 19: 135–45.doi:10.1016/S0747-5632(02)00058-4

Carpenter, S., H. Pashler, J. Wixted, and E. Vul. 2008. ‘‘TheEffects of Tests on Learning and Forgetting.’’ Memory andCognition 36: 438–48. doi:10.3758/MC.36.2.438

Casey, M.B. 1996. ‘‘Understanding Individual Differences inSpatial Ability within Females: A Nature/Nurture Interac-tionist Framework.’’ Developmental Review 16: 241–60.doi:10.1006/drev.1996.0009

Coolican, J., and M. Peters. 2003. ‘‘Sexual Dimorphism inthe 2D/4D Ratio and Its Relation to Mental Rotation Per-formance.’’ Evolution and Human Behavior 24: 179–83.doi:10.1016/S1090-5138(03)00010-2

Cowan, N., and C.C. Morey. 2006. ‘‘Visual Working MemoryDepends on Attentional Filtering.’’ Trends in CognitiveSciences 10/4: 139–41. doi:10.1016/j.tics.2006.02.001

Csatho, A., A. Osvath, K. Karadi, E. Bicsak, J. Manning, andJ. Kallai. 2003. ‘‘Spatial Navigation Related to the Ratioof Second to Fourth Digit Length in Women.’’ Learning andIndividual Differences 13: 239–49.doi:10.1016/S1041-6080(02)00093-6

Dabbs, J.M., Jr., E. Chang, R.A. Strong, and R. Milun. 1998.‘‘Spatial Ability, Navigation Strategy, and GeographicKnowledge among Men and Women.’’ Evolution and HumanBehavior 19: 89–98. doi:10.1016/S1090-5138(97)00107-4

Falter, C., M. Arroyo, and G. Davis. 2006. ‘‘Testosterone:Activation or Organization of Spatial Cognition?’’ BiologicalPsychology 73: 132–40.doi:10.1016/j.biopsycho.2006.01.011

Fenner, J., D. Heathcote, and J. Jerrams-Smith. 2000. ‘‘TheDevelopment of Wayfinding Competency: AsymmetricalEffects of Visuo-spatial and Verbal Ability.’’ Journal of Envi-ronmental Psychology 20: 165–75.

Fischer, M.H. 2001. ‘‘Probing Spatial Working Memory withthe Corsi Blocks Task.’’ Brain and Cognition 45: 143–54.doi:10.1006/brcg.2000.1221

Focquaert, F., M.S. Steven, G.L. Wolford, A. Colden, and M.S.Gazzaniga. 2007. ‘‘Empathizing and Systemizing CognitiveTraits in the Sciences and Humanities.’’ Personality andIndividual Differences 43: 619–25.doi:10.1016/j.paid.2007.01.004

Gouchie, C., and D. Kimura. 1991. ‘‘The Relationship betweenTestosterone Levels and Cognitive Ability Patterns.’’ Psycho-neuroendocrinology 16: 323–34.doi:10.1016/0306-4530(91)90018-O

Graf, S., T. Lin, and Kinshuk. 2008. ‘‘The Relationshipbetween Learning Styles and Cognitive Traits – GettingAdditional Information for Improving Student Modelling.’’Computers in Human Behavior 24: 122–37.doi:10.1016/j.chb.2007.01.004

Halpern, D.F. 2000. Sex Differences in Cognitive Abilities.Mahwah, NJ: Lawrence Erlbaum.

Harrower, M. 2007. ‘‘The Cognitive Limits of AnimatedMaps.’’ Cartographica 42: 349–57.doi:10.3138/carto.42.4.349

Hardwick, S.W., L.L. Bean, K.A. Alexander, and F.M. Shelley.2000. ‘‘Gender vs. Sex Differences: Factors Affecting Perfor-mance in Geographic Education.’’ Journal of Geography 99/6: 238–44. doi:10.1080/00221340008978974

Holsanova, J., N. Holmberg, and K. Holmqvist. 2009. ‘‘Read-ing Information Graphics: The Role of Spatial Contiguityand Dual Attentional Guidance.’’ Applied Cognitive Psychol-ogy 23: 1215–26. doi:10.1002/acp.1525

Jager, G., and A. Postma. 2003. ‘‘On the Hemispheric Spe-cialization for Categorical and Coordinate Spatial Relations:A Review of the Current Evidence.’’ Neuropsychologia 41:504–15. doi:10.1016/S0028-3932(02)00086-6

Janowsky, J. 2006. ‘‘Thinking with Your Gonads: Testosteroneand Cognition.’’ Trends in Cognitive Sciences 10: 77–82.doi:10.1016/j.tics.2005.12.010

Kimura, D. 1989. ‘‘How Sex Hormones Boost or Cut Intellec-tual Ability.’’ Psychology Today 23: 62–66.

———. 2000. Sex and Cognition. Cambridge, MA: MIT Press.

Kosslyn, S.M., W.L. Thompson, D.R. Gitelman, and N.M.Alpert. 1998. ‘‘Neural Systems That Encode Categoricalversus Coordinate Spatial Relations: PET Investigations.’’Psychobiology 26: 333–47.

Kulhavy, R.W., W.A. Stock, and W.A. Kealy. 1993. ‘‘HowGeographic Maps Increase Recall of Instructional Text.’’ Edu-cational Technology Research and Development 41: 47–62.doi:10.1007/BF02297511

Lloyd, R.E., and R.L. Bunch. 2003. ‘‘Technology and Map-Learning: Users, Methods, and Symbols.’’ Annals of the



Association of American Geographers 93: 828–50.doi:10.1111/j.1467-8306.2003.09304004.x

———. 2005. ‘‘Individual Differences in Map Reading SpatialAbilities Using Perceptual and Memory Processes.’’ Carto-graphy and Geographic Information Science 32: 33–46.doi:10.1559/1523040053270774

———. 2008. ‘‘Explaining Map-Reading Performance Effi-ciency: Gender, Memory, and Geographic Information.’’Cartography and Geographic Information Science 35: 171–202. doi:10.1559/152304008784864677

Lloyd, R.E., M.E. Hodgson, and A. Stokes. 2002. ‘‘Visual Cate-gorization with Aerial Photographs.’’ Annals of the Associa-tion of American Geographers 92: 241–66. doi:10.1111/1467-8306.00289

Loehlin, J.C., D. McFadden, S.E. Medland, and N.G. Martin.2006. ‘‘Population Differences in Finger-Length Ratios: Ethnic-ity or Latitude?’’ Archives of Sexual Behavior 35: 739–42.doi:10.1007/s10508-006-9039-1

Manning, J. 2002. Digit Ratio: A Pointer to Fertility, Behaviorand Health. New Brunswick, NJ: Rutgers University Press.

Manning, J., D. Scutt, J. Wilson, and D. Lewis-Jones. 1998.‘‘The Ratio of 2nd to 4th Digit Length: A Predictor of SpermNumber and Concentrations of Testosterone, LuteinizingHormone, and Oestrogen.’’ Human Reproduction 13: 3000–3004. doi:10.1093/humrep/13.11.3000

Mayer, R., and R. Moreno. 2003. ‘‘Nine Ways to Reduce Cog-nitive Load in Multimedia Learning.’’ Educational Psycholo-gists 38: 43–52. doi:10.1207/S15326985EP3801_6

McDaniel, M., H. Roediger, and K. McDermott. 2007. ‘‘Gen-eralizing Test-Enhanced Learning from the Laboratory tothe Classroom.’’ Psychonomic Bulletin and Review 14: 200–206.

Nettle, D. 2007. ‘‘Empathizing and Systemizing: What AreThey, and What Do They Contribute to Our Understandingof Psychological Sex Differences?’’ British Journal of Psy-chology 98: 237–55. doi:10.1348/000712606X117612

Ozuru, Y., K. Dempsey, and D. McNamara. 2009. ‘‘PriorKnowledge, Reading Skill, and Text Cohesion in the Com-prehension of Science Texts.’’ Learning and Instruction 19:228–42. doi:10.1016/j.learninstruc.2008.04.003

Paas, F., and J. van Merrienboer. 1993. ‘‘The Efficiency ofInstructional Conditions: An Approach to Combine Mental-Effort and Performance Measures.’’ Human Factors 35: 737–43.

Patton, D.K., and Lloyd, R.E. 2009. ‘‘Asymmetrical Learning ofLocations on Maps: Implicit Learning, Prior Knowledge andSex Differences.’’ Cartographic Perspectives 63: 4–34.

Roediger, H., and J. Karpicke, J. 2006. ‘‘Test-Enhanced Learn-ing: Taking Memory Tests Improves Long-Term Retention.’’Psychological Science 17: 249–55.doi:10.1111/j.1467-9280.2006.01693.x

Sanders, G., M. Sjodin, and M. Chastelaine. 2002. ‘‘On theElusive Nature of Sex Differences in Cognition: HormonalInfluences Contributing to Within-Sex Variation.’’ Archivesof Sexual Behavior 31: 145–52.

Saucier, D., S. Green, J. Leason, A. MacFadden, S. Bell, andL. Elias. 2002. ‘‘Are Sex Differences in Navigation Causedby Sexually Dimorphic Strategies or by Differences in theAbility to Use the Strategies?’’ Behavioral Neuroscience116: 403–10. doi:10.1037/0735-7044.116.3.403

Seufert, T. 2003. ‘‘Supporting Coherence Formation in Learn-ing from Multiple Representations.’’ Learning and Instruc-tion 13: 227–37. doi:10.1016/S0959-4752(02)00022-1

Signorella, M., and W. Jamison. 1986. ‘‘Masculinity, Femi-ninity, Androgyny, and Cognitive Performance: A Meta-analysis.’’ Psychological Bulletin 100: 207–28. doi:10.1037/0033-2909.100.2.207

Slotnick, S., and L. Moo. 2006. ‘‘Prefrontal Cortex Hemi-spheric Specialization for Categorical and Coordinate VisualSpatial Memory.’’ Neuropsychologia 44: 1560–68.doi:10.1016/j.neuropsychologia.2006.01.018

Sommer, T., E. Schoell, and C. Buchel. 2008. ‘‘AssociativeSymmetry of the Memory for Object-Location Associationsas Revealed by the Testing Effect.’’ Acta Psychologica 128:238–48. doi:10.1016/j.actpsy.2008.01.003

Sweller, J. 1988. ‘‘Cognitive Load during Problem Solving:Effects on Learning.’’ Cognitive Science 12: 257–85.doi:10.1207/s15516709cog1202_4

van der Lubbe, R., M. Scholvinck, J. Kenemans, and A.Postma. 2006. ‘‘Divergence of Categorical and CoordinateSpatial Processing Assessed with ERPs.’’ Neuropsychologia44: 1547–59. doi:10.1016/j.neuropsychologia.2006.01.019

Verdi, M.P., and R.W. Kulhavy. 2002. ‘‘Learning with Mapsand Texts: An Overview.’’ Educational Psychology Review14: 27–46.

Voyer, D., A. Postma, B. Brake, and J. Imperato-McGinley.2007. ‘‘Gender Differences in Object Location Memory: AMetaanalysis.’’ Psychonomic Bulletin and Review 14: 23–38.

Wilson, D.E., C.M. MacLeod, and M. Muroi. 2008. ‘‘Practice inVisual Search Produces Decreased Capacity Demands ButIncreased Distraction.’’ Perception and Psychophysics 70:1130–37. doi:10.3758/PP.70.6.1130

Yarrow, K., P. Brown, and J.W. Krakauer. 2009. ‘‘Inside theBrain of an Elite Athlete: The Neural Processes That SupportHigh Achievement in Sports.’’ Nature Reviews Neuroscience10: 585–96. doi:10.1038/nrn2672

Zinser, O., D.L. Palmer, and C.R. Miller. 2004. ‘‘Site Distance,Gender, and Knowledge of Geographic Sites.’’ Sex Roles 51:661–86. doi:10.1007/s11199-004-0717-y

(Continued)



Appendix: The Text Supplying Physical, Cultural, andTravel Information about the Isle of Dogs

Physical. The Isle of Dogs is located between the Atlantic

Ocean and the North Sea. The island got its name from

the canines left on the island by passing Viking raiders.

Rivers flow from the central highlands to the sea in three

of the provinces. The English established the first perma-

nent settlement at the mouth of the Annan River in 1022

AD. The climate is warm in the summer and cold in the

winter. Storms from the Atlantic Ocean provide frequent

and high levels of precipitation.

Cultural. The modern island has four provinces. Nussex

has the smallest population and 3 counties and Essex has

the largest population and 5 counties. Trent, the capital

city, is centrally located in Essex and is the main seat of

government. Other major cities are Kenly in Sussex that

supports a thriving fishing industry, Wells in Wessex that

excels in manufacturing and Alban in Nussex, that is the

home of a growing recreation trade.

Travel. The island has a simple highway system that

connects the towns and cities. There is an excellent and

inexpensive bus system serving the entire island. Cars can

only be rented at the airport. Travelers coming by ship

generally come through the port of Wells while those

arriving by air land at the airport near Trent. Travelers

from outside of Europe will need to show a passport.

Hotel rates are highest in Trent and less expensive in

other locations.



Copyright of Cartographica is the property of University of Toronto Press and its content may not be copied or

emailed to multiple sites or posted to a listserv without the copyright holder's express written permission.

However, users may print, download, or email articles for individual use.

learning geographic information from a map and text: learning

Documents