Introduction
The rapidly evolving field of online mapping offers a diverse range of techniques, technologies, and data sources, catering to both casual and professional users Increased internet access and interactive mapping tools have sparked interest among educators in incorporating these resources into learning activities However, there is a notable lack of research focused on the use of interactive online maps in educational contexts, particularly those that integrate multiple thematic layers within a single map.
This study investigates how the number of layers on a web-based map influences middle school students' accuracy and response times when answering geographic questions It also explores the impact of hillshade on students' performance in terms of response time and accuracy, even for questions that do not necessitate terrain information.
From 2006 to 2008, the Oregon Geographic Alliance partnered with middle school teachers and Portland State University students to develop the Student Atlas of Oregon This atlas features a variety of thematic maps and visual geographic concepts designed to enhance the curriculum for upper elementary and middle school students in Oregon The topics for the maps were carefully chosen, drawing inspiration from other child-friendly atlases and contemporary examples.
2 curricula, then further reduced to a collection that best met the needs of the widest group of students
The traditional Atlas was limited to static maps in print or PDF format, but feedback from collaborators indicates a strong interest in developing an interactive online version This new format would enable students to toggle data layers on and off, enhancing their learning experience The capability to incorporate additional data layers into interactive maps has gained popularity among online mapping tools, benefiting both students and instructors beyond the classroom setting.
Before creating a map, it is essential to identify suitable content for upper elementary and middle school students and to understand the map skills they are expected to possess This research was guided by educational content standards, which informed the selection of both maps and relevant content.
In Oregon, the Department of Education (ODE) sets curriculum guidelines and learning benchmarks for standardized content, although these are not mandatory and local school districts have the final say on instruction Despite this flexibility, ODE's content standards and curriculum goals provide a useful framework for this study, particularly as they outline expectations for 5th graders to "examine and understand how to prepare maps, charts, and other visual representations to locate places and interpret geographic information."
(ODE 2005, 55) This standard is exemplified in the grade level learning targets by using "maps to determine population trends, precipitation, temperature and ethnic distribution" (ODE 2003, 15)
Thematic maps and general reference maps are the two primary types of maps utilized today While reference maps emphasize the identification of various geographic features, both natural and human-made, their main objective is to convey precise locations.
1996) As adults, we may frequently encounter a reference map in the form of a topographic map, such as a USGS Quad sheet
Thematic maps concentrate on a specific topic or "theme," presenting either quantitative or qualitative data Qualitative thematic maps illustrate the distribution of various phenomena, such as the locations of downhill ski areas.
Figure 1: A Qualitative Map This type of map is often used to show where a certain phenomenon occurs, such as downhill ski areas in this example
Quantitative thematic maps serve to convey the extent of specific activities or phenomena in various locations, providing a geographic context to tabular data, such as rainfall measurements at different sites (Dent, 1996).
Figure 2: A Quantitative Thematic Map This map uses area symbols to represent the amount of precipitation that parts of the state receive each year
The Student Atlas is largely comprised of thematic maps, so it is important to discuss this type of map and the potential for making thematic maps interactive
Creating a thematic map involves overlaying a base map with a specific thematic layer, ensuring that the base map contains only essential geographic information for user comprehension (Dent, 1996) With the rise of computer-based cartography, map makers can swiftly select and combine data for both base and thematic layers Interactive layer maps empower users, allowing them to manipulate and explore multiple data sources, similar to the capabilities of map creators.
Combining multiple data layers allows map makers to create a comprehensive map, while interactive maps enable users to select the specific layers they want to include in their final visualization.
Symbology is the method of using marks or symbols to convey information on a map, and map makers must select these symbols carefully based on the information's nature and the map's scale In small-scale maps, points represent specific locations like cities, while lines depict linear features such as roads and rivers Area symbols illustrate two-dimensional phenomena, such as lakes or larger features like cities Research on symbol selection, including considerations for children, is limited but important for effective map communication.
1982, Wiegand 2006, Michaelidou 2004, Gaspers 2007) Appendix A includes a summary of symbology guidelines
Literature Review
Understanding the intended users and the media through which they will view a map is crucial for effective map creation This ensures the application of appropriate cartographic techniques tailored to communicate effectively with users In this thesis, the focus is on children as the primary audience and computer screens as the medium Therefore, the literature review is organized into two key sections: one addressing map usage by children and the other focusing on the development and utilization of web-based and interactive maps.
Research on children's use of maps has evolved significantly since the 1960s and 70s, initially emphasizing their developmental stages and understanding of geographic symbolism In the past two decades, the focus has shifted towards involving children in the map-making process and fostering collaboration among students Recent studies have explored specific applications of maps in enhancing student learning This article reviews inquiries into children's engagement with maps and atlases.
To enhance geographic learning, effective communication of geographic relationships through suitable maps is essential Research on creating thematic maps for children primarily emphasizes the importance of choosing appropriate symbology and managing map complexity.
In 1965, Barbara Bartz conducted a study on over 200 students aged 9-14 to investigate their interaction with maps and interpretation of symbology Her research led to key recommendations for designing maps for children, including labeling keys with symbols of the same size as those used on the map, and incorporating obvious items such as water Bartz also suggested strategic type and labeling techniques, such as placing labels to reinforce geographic locations and exaggerating font size differences to classify features These findings have significantly influenced best design practices for children's maps, with further recommendations summarized in Appendix A.
Jack Miller developed a set of guidelines for classroom maps based on his research with middle school students in 1982 He provided students with one of four distinct maps, each designed with different elements, to determine which was most effective in helping them answer questions However, neither Miller nor Bartz applied their guidelines in a practical retest with students, and their recommendations primarily cater to adult audiences Dent (1996) highlights the significance of type placement in enhancing the visibility of geographic features, while Bartz (1965) stresses that this practice is particularly crucial for children.
Elementary and middle school students are increasingly exposed to thematic maps, which research indicates can help them comprehend geographic relationships effectively.
In a 2007 study, Karen Trifonoff engaged 2nd grade students with computer-based thematic maps to explore differences in the quantity and distribution of map objects, followed by timed responses and discussion questions to assess comprehension Similarly, Stephanie Gaspers investigated the effectiveness of three types of thematic maps in helping students understand population amounts and distributions Both researchers concluded that younger students can effectively utilize qualitative and quantitative thematic maps to process geographic information.
Michaelidou et al (2004) investigated the use of various textbook maps by elementary school students, focusing on grades 3-6, and analyzed their ability to respond to questions based on the maps' complexity The study highlighted that maps differed in scale, complexity, thematic layers, and terrain representation The authors criticized the common practice of simplifying maps to a minimal number of thematic layers, arguing that this approach renders them too abstract and less useful Their research assessed students' skills in identifying four geographic relationships: locations, types of geographic attributes, distribution of attributes, and connections between locations with similar attributes.
In a study utilizing nearly identical maps, one map featured fewer thematic layers by omitting certain building types While performance differences between schools and genders were minimal, variations were noted across grade levels Students using maps with multiple thematic layers performed better overall, except for fourth graders, who excelled with maps containing fewer layers Conversely, the reduced-layer maps negatively impacted older students in grades 5 and 6 when it came to extracting spatial relationships.
Research by Michaelidou et al (2004) indicates that students performed better on political maps compared to physical maps, likely because political maps present simpler background information The absence of topographical and landform details in political maps eliminates competing visual elements, enhancing clarity Consequently, the authors suggest that the trend of simplifying maps in elementary textbooks by removing thematic layers should be re-evaluated.
Landform representation is crucial for providing geographical context, yet it presents challenges in map readability Peter Collier highlights that terrain portrayal often fails to convey relief effectively, which can hinder the clarity of other map information Research by Michaelidou et al indicates that shaded relief is particularly beneficial for middle school students in understanding slope and elevation However, other studies suggest that hillshades may complicate the extraction of non-landform information from maps.
In a study focusing on general reference maps, participants were tasked with locating symbols, while the relationship between terrain and other mapped features was overlooked Sharon Muir (1985) discovered that many elementary students confused colors in hypsometric relief shading for vegetation rather than elevation Meanwhile, Patrick Wiegand (2006) highlighted the potential of pseudo-3D terrain displays using Digital Elevation Models (DEMs), although their effectiveness can be influenced by students' prior experiences with hills and slopes.
Interactive maps empower users to take on the role of map makers, enhancing their geographic education Research by Leinhardt, Stainton, and Bausmith (1998) highlights the advantages of engaging seventh-grade students in map construction, as it fosters a deeper understanding of map content and the implicit decisions involved in map-making.
Bausmith and Leinhardt's study examines the National Geography Standards in education, focusing on the representation of layers in maps The research involved students working individually or collaboratively to recreate maps at varying scales.
Students' success in the map enlargement task was significantly influenced by their ability to recognize the interconnections between map elements As additional layers were incorporated into the map, the relationships between these elements became increasingly complex, highlighting the importance of understanding these connections for effective map interpretation.
12 occurred, and more chances for inaccuracies to arise developed With each added layer, it became more difficult to consider the many interconnections between elements"
Methodology
This research aims to explore the impact of thematic layer maps on middle school students' responses and response times to geographic questions By utilizing a web-based map interface, the study enables simultaneous engagement of a large student group and streamlines data collection The methodology encompasses several stages, including map design, question formulation, user testing, pilot testing, data collection, and analysis.
Cartographic design practices for static print maps and online maps share many similarities, as highlighted by Kraak (2001), while also considering the unique needs of children’s maps (Miller 1984, Wiegand 2006) Online maps employ design techniques akin to those used in print, but cartographers must be mindful of the output medium, such as computer screens or projectors, which can influence the final appearance The student atlas project utilized digital data sources to create print maps, demonstrating that this data can also be effectively transformed into online maps with minimal adjustments However, when designing dynamic online maps, careful consideration of complexity and resolution is essential to enhance user experience.
An online map often features competing visual elements such as menus, toolbars, and other web pages, which can distract users and hinder their ability to navigate the map effectively To enhance usability, it is essential to maintain a low graphic and information density, as suggested by Kraak (2001) This involves ensuring adequate white space or padding between elements on the map and its interface, ultimately simplifying the user experience.
This guideline supports Leinhardt and Bausmith’s advice that student maps should feature multiple layers with minimal data in each layer Consequently, the data for the print atlas has been simplified to match the online map's resolution, which involved smoothing the edges of larger features like counties and omitting smaller elements such as islands Additionally, the process of selection and combination further exemplifies generalization techniques that decrease the information presented on the map while maintaining accuracy at this scale.
Figure 6: A web map includes many competing visual elements
Figure 8: Example of selection and combination Some small polygons were removed from the original
(selection) and some were merged to reduce the number of smaller features (combination)
Online maps present unique challenges compared to traditional print maps, as they can be displayed on various media, including CRT monitors, TFT flat-panel displays, LCD projectors, and printed formats Each type of display has its own characteristics, requiring careful consideration in design and usability.
Figure 7: Example of simplification Lines were smoothed by removing additional points from the vector file and islands were removed because they were unimportant at the scale being displayed
For effective web design, especially for children's use, it's advisable to utilize a conservative color palette, such as a "web-safe" selection of colors that are display-friendly (Feringa 2001, Van der Worms 2001) Despite the trend of downplaying the web-safe palette's relevance (Weinman 2008), children benefit from simple, high-contrast color schemes (Miller 1982) with vibrant saturation (Buckingham and Harrower 2007) Additionally, ensuring that maps are designed to fit a standard 1024 x 768 pixel display is essential, as this resolution is commonly used in computer labs.
Many print versions of student atlases often omit neighboring states, but the map in question effectively includes outlines and labels for California, Washington, Idaho, and the Pacific Ocean, providing essential context as recommended by Patrick Wiegand (2006).
Effectively conveying terrain information is crucial, yet it presents challenges Although hillshading can hinder map users' comprehension of thematic data (DeLucia 1972), research indicates that this technique is particularly beneficial for middle school students, as it helps illustrate elevation and facilitates comparisons regarding elevation (Michaelidou et al 2004).
Due to time constraints and technical limitations, I was unable to develop a fully interactive layer map that would engage students with questions and record their responses Instead, I created a series of individual maps featuring various layers to simulate an interactive experience While each map may include design elements that deviate from traditional mapping conventions, these choices were intentionally made to mimic the functionality of an interactive layer map For instance, the placement of county seat labels might appear inappropriate with only two layers visible, but they were strategically positioned to enhance interaction with labels from other layers.
26 may not be visible I decided to keep the design the same on each layer map for consistency between various layer maps
To create effective questions for middle school students, it is essential to align them with state curriculum guidelines and the National Geography Standards According to these standards, students should be able to utilize thematic maps to address inquiries regarding human distributions, employ labels and symbols to identify physical and human features, and analyze the locations of places to understand why specific sites are chosen for various activities.
Meeting all these criteria in a small set of questions is challenging, though good examples can be found in the State of Oregon Department of Education’s Social
The Science Benchmark 3 sample test features two questions linked to a map depicting early agricultural sites in Europe, which includes outlines of the continent, rivers, and designated agricultural locations The first question explores the relationship between the proximity of these sites and various landforms, while the second question examines population movement in relation to the age and location of the agricultural sites Together, these questions align with National Geography Standards 1 and 2, utilizing a single, informative map.
My test included multiple choice questions that were paired with a map
Questions 1 to 5 focused on a specific precipitation layer, while Questions 6 and 7 utilized agricultural product layers The students were tasked with identifying locations on the map in Questions 3 and 4.
27 location on the map and identify how much precipitation occurs at that location
In this analysis, we explore the optimal locations for specific agricultural activities by examining the correlation between precipitation levels and crop suitability By identifying counties with significant agricultural outputs, we can assess the impact of these climatic factors on farming practices Furthermore, we compare two distinct thematic layers to highlight variations in agricultural productivity and resource allocation across different regions This comprehensive approach facilitates a deeper understanding of how environmental conditions influence agricultural success.
In the study, three questions (7, 8, and 9) included a brief written follow-up, while various biographical questions were posed to identify differences based on age, gender, and computer exposure An example of the map interface and several related questions can be found in Appendix D.
This research involved students answering questions similar to those in a typical classroom setting, utilizing maps with various thematic layers Each student received a single map, with the number of layers differing among them The study recorded accuracy scores and response times to analyze the relationship between student performance and the number of layers present on the maps To simplify the process and minimize the impact of interactive features, static maps were used instead of dynamic layer maps.
I found several teachers willing to participate through the Oregon Geographic Alliance and word of mouth An informational flyer used to recruit potential volunteers can be found in Appendix C
Results
This chapter summarizes the collected data and analyzes the results, which are organized by question Each question utilized a unique set of maps, and a different map was randomly assigned to the students for each inquiry.
Sections 10 and 11 provide a comprehensive summary of the maps utilized for each question, detailing the quantity and types of layers present in each map, along with the number of sampled users associated with them.
Questions 8 and 9 are not included because those questions were not map reading tasks and did not evaluate response time and score
The following null hypotheses were used when appropriate for each question when conducting statistical tests:
H 01 : There is no significant relationship between the number of layers on a map and student scores
H 02 : There is no significant relationship between the number of layers on a map and student response time
H o3 : There is no significant difference between scores for students who had a map with a hillshade and those who did not
H o4 : There is no significant difference between response times for students who had a map with a hillshade and those who did not
Figure 10: Map distribution summary There were a total of 13 maps used in Questions 1-4, ranging from
The study involved 2 to 7 layers, with the total sample size for each question indicated by the numbers above the arrows Sample sizes varied from a minimum of 14 to a maximum of 50 A comprehensive list of the layers utilized in each map is provided on the right.
The distribution summary of maps, as illustrated in Figure 11, includes a total of 13 maps utilized for Questions 5-7, featuring between 2 to 7 layers each The figures above the arrows represent the overall sample size for each question, along with the sample count for each specific map Additionally, the layers incorporated in each map are clearly labeled to the right of the respective maps.
To conduct a thorough analysis, I first assessed the overall results and examined the impact of biographical factors on these outcomes During the data collection phase, students provided information on computer ownership, perceived difficulty of the exercise, as well as their age and gender I analyzed the overall scores and response times for map reading questions (Questions 1-7) in relation to age, gender, school, and computer screen size, summarizing the findings in Table 2 However, I did not receive sufficient responses regarding home computer access to conduct a meaningful analysis.
Table 2: Analysis of biographical traits on overall time spent and score for the map-based questions
Comparison (Population variable and factor) Test Test Result
Age and Score (df=3) Kruskal-Wallis 4.83 (χ2) 7.82 No
Gender and Score Mann-Whitney U -0.23 (Z) -1.65 No
Gender and Time Mann-Whitney U -1.12 (Z) -1.65 No
School and Score (df=4) Kruskal-Wallis 1.38 (χ2) 9.49 No
Screen size and score Mann-Whitney U -0.29 (Z) -1.65 No
Screen size and time Mann-Whitney U -0.01 (Z) -1.65 No
*The Question 13 (“Do you have a computer at home”) did not have enough responses to perform any meaningful statistical test
The analysis revealed that age and school were the only biographical factors significantly impacting overall test results, specifically in terms of response time It is plausible that these two factors, age and classroom environment, are interconnected.
In a study analyzing response times among 7th graders, it was found that comparing results across different classrooms for 13-year-olds was challenging due to potential site-specific factors, such as network latency and teacher influence The Kruskal-Wallis test revealed no significant differences in response times for 10-12-year-olds, highlighting that age was a crucial factor for 13-year-olds When excluding the 7th-grade classroom from the evaluation, significant results persisted, suggesting that classroom dynamics might play a role, though insufficient data prevented definitive conclusions Notably, while there was a significant difference in reading speed among 7th graders compared to 5th and 6th graders, this did not translate to improved mastery of the material, indicating that faster reading does not equate to better understanding.
These results suggest that age may have an impact on the overall results
The random distribution of maps to students diminished the influence of age on test outcomes, as the quicker response times of older students were spread across various questions Consequently, this randomization reduced the likelihood of age affecting the results linked to any specific map.
In Table 3, I have provided a key to the results for each question, highlighting the comparisons between independent variables (Layers, Hillshade) and dependent variables (Score, Response Time) The table also includes page numbers for the test results, with significant findings emphasized for clarity.
Table 3: Summary of Statistical Findings Page numbers are listed for corresponding test results Highlighting indicates that the test results were significant
Layer Hillshade Layer Hillshade Layer Hillshade Layer Hillshade Layer Hillshade Layer Hillshade
Results
Question 1 (Figure 12) served as an introductory exercise designed to familiarize students with map interaction It marked the first instance where students independently engaged with the map to answer a question.
Figure 12: Question 1 and a sample map Students were randomly given one of three maps, each with a differing number of layers
Students were tasked with using a key to compare regional precipitation through multiple-choice answers, utilizing a map and its accompanying key Three alternative questions were provided, varying in complexity with options of two, three, and five layers Notably, the five-layer map featured a hillshade layer, enhancing the depth of the analysis.
Look at the map above to answer the following question: Which of the following regions gets the most precipitation?
In my initial analysis, I explored the correlation between the number of layers on a map and students' scores The findings, detailed in Table 4, reveal minimal variation between the number of map layers and the average score achieved by the students.
Table 4: Descriptive Statistics for Question 1, layer evaluation
Layers Count Percent Correct Median Time
The Pearson’s Chi-Square test is typically preferred for assessing the relationship between expected and observed results; however, it necessitates that each expected category has a count of five or more, with fewer than 20% of categories having counts below five and all counts exceeding one (McGrew and Monroe 2000, 156) Unfortunately, my data did not meet these criteria, prompting the use of SPSS’s Exact test, a proprietary method suitable for small data sets that fail to satisfy the Pearson’s requirements (IBM 2010) The findings from the Exact test revealed no significance, leading me to accept the null hypothesis.
The descriptive statistics presented in Table 4 reveal variations in mean response times across different maps As depicted in Figure 13, there is a noticeable increase in mean response time correlated with the number of layers To assess whether these differences in mean response times were statistically significant and not due to random chance, the Kruskal-Wallis test was employed.
The Kruskal-Wallis test yielded a Chi-Square result of 26.50, surpassing the critical value of 5.99 at p = 0.05 with two degrees of freedom Consequently, the null hypothesis was rejected in favor of the alternative hypothesis, indicating a significant relationship between the number of layers on a map and student response time.
Next I investigated the effect that a hillshade has on student response time and score I collected the descriptive statistics for the second part of this question in Table 5
Table 5: Descriptive Statistics for Question 1, hillshade evaluation
Map Count Percent Correct Median Time
I used the Exact test to evaluate the differences between the scores on maps with and without hillshades because the requirements for the Pearson’s Chi-Square test were
Figure 13: Response times for each map in Question 1.
46 not met The Exact test results did not indicate any significance, so I accepted the null hypothesis
The Mann-Whitney U test was employed to assess the relationship between the presence of hillshade on maps and student response times The findings, as shown in Table 5, reveal a significant difference in mean response times between maps with hillshade and those without Figure 14 illustrates a box plot comparing response times, highlighting notable variations in both mean and range The test yielded a Z score of -3.85, surpassing the critical value of 1.96 at p < 0.05, leading to the rejection of the null hypothesis, which posited no relationship between hillshade presence and response time.
Figure 14: Response times for the maps with and without hillshades for Question 1.
Question 2 closely resembles the first, featuring maps with four, six, and seven layers Students received one of these four maps and were tasked with analyzing it.
In the analysis of precipitation across various regions, two maps were utilized, each featuring six distinct layers, with one map including a hillshade effect while the other did not To summarize the findings, a table of descriptive statistics (Table 6) was compiled for the initial section, highlighting the regions that received the most precipitation.
Table 6: Descriptive Statistics for Question 2, Layer evaluation
Layers Count Percent Correct Median Time
The data failed to satisfy the criteria for the Pearson Chi-Square test, prompting the use of Exact analysis in SPSS, which also produced non-significant results Consequently, the null hypothesis was accepted.
The distribution of response times, categorized by the number of layers on the map, is illustrated in Figure 15 To analyze these response times across different maps, the Kruskal-Wallis test was employed, resulting in a Chi-Square value of 8.65 with two degrees of freedom and a significant p-value.
The analysis yielded a significant Chi-Square value of 9.15 with three degrees of freedom and a p-value of 0.05, indicating a notable relationship between the number of layers on a map and student response time By treating the two distinct maps as separate categories, the results consistently supported the rejection of the null hypothesis.
Figure 15: Response times for different maps in Question 2 The map labeled 6C did not include a hillshade as one of the layers
In my study, I examined the impact of hillshade on student response times and scores Utilizing the Exact test to assess the correlation between student scores and the presence of hillshade, I found that the data did not satisfy the criteria for the Pearson Chi-Square test Ultimately, the results indicated no significant relationship, leading to the acceptance of the null hypothesis.
A visual comparison of the relationship between response times and presence of hillshade (Figure 16) does not reveal any obvious difference
Figure 16: Response times for maps with and without hillshades in Question 2
In assessing the impact of hillshade on response time, I conducted a Mann-Whitney U test, yielding a Z score of -0.12 This result is not statistically significant at the p = 0.05 level, leading to the acceptance of the null hypothesis.
and 4
Questions 3 and 4 were selected from a shared question pool, allowing for random assignment of two questions to each student This method aimed to gather a substantial sample of similar questions while ensuring that each individual map received sufficient responses for statistical analysis Ultimately, this approach yielded my largest sample size.
In Questions 3 and 4, students were tasked with identifying a specific city in Oregon and reporting the expected precipitation levels for that location Figure 17 illustrates a map featuring two key layers: the county seats and the precipitation data.
50 information, which was all that was needed to answer the question The map with seven layers contained the most competing visual elements, including a hillshade
Figure 17: Sample question and map used in Questions 3 and 4 on the test
As the number of layers on the maps increased, I anticipated a rise in response time and a decline in the mean score To illustrate these findings, I compiled the descriptive statistics for Questions 3 and 4 in Table 7.
Table 7: Summary statistics for Questions 3 and 4
Layers Count Percent Correct Median Time
Find Gold Beach on the map It is on the southwest coast About how many inches of precipitation does it get?
I conducted a Pearson’s Chi-Square test to explore the relationship between the number of layers and student scores, utilizing a sufficient dataset that met the test's requirements The Chi-Square value obtained was 21.837, surpassing the critical value of 11.07 for degrees of freedom (df) equal to 5 at a significance level of p = 0.05 Consequently, I can reject the null hypothesis and accept the alternative hypothesis, indicating a significant relationship between the number of layers and student scores.
In this study, I examined the correlation between response time and the number of layers using the Kruskal-Wallis test The findings, illustrated in Figure 18, indicated that the critical value for degrees of freedom (df) equal to 5 at a significance level of p = 0.05 is 11.07 However, the test result of 10.68 did not reach significance, leading to the acceptance of the null hypothesis.
My research examined how hillshade affects response time and scores The descriptive statistics presented in Table 8 reveal only a slight difference in the average scores between maps that include hillshade and those that do not.
Table 8: Descriptive Statistics for Question 3, Hillshade evaluation
Map Count Percent Correct Median Time
Figure 18: Response times for each map in Questions 3 and 4.
I conducted a Pearson Chi-Square test to assess the correlation between scores and the presence of hillshade The analysis yielded a Chi-Square value of 1.34, which was not statistically significant, with 3 degrees of freedom and a p-value of 0.05 Consequently, I accepted the null hypothesis.
The Kruskal-Wallis test was employed to assess the response times of students using maps with hillshades compared to those without The results indicated no significant visual difference in response times between the two groups, similar to the findings in Question 2.
Figure 19: Response times for maps with and without a hillshade in Questions 3 and 4 The results are nearly identical
The results of the Kruskal-Wallis test gave a Chi-Square value of 0.421, which is insignificant with df = 1 and p= 0.05 so I accepted the null hypothesis.
Question 5 presented a story problem where students had to select the most suitable region for wheat cultivation from a list of options Accompanying this question was Figure 20, which illustrated one of six distinct maps, with each student being randomly assigned one of these maps for their evaluation.
Figure 20: Question 5 asked students to choose the best region to start a wheat farm
In instances where multiple correct answers were possible, I awarded full points to any region deemed relatively drier than the other options For example, both the "Northern Basin and Range" and the "Blue Mountains" received full points, as they are significantly drier compared to the "Cascades" and "Coast Range" based on precipitation data.
In my investigation of the correlation between the number of layers on the map and student performance, I analyzed both scores and response times The findings, detailed in Table 9, present the descriptive statistics for each map categorized by the number of layers.
As a wheat farmer from Montana considering a new venture in Oregon, it’s essential to choose a region with a suitable climate for wheat cultivation Opting for areas in Eastern Oregon, such as the Columbia Basin or the High Desert, would be ideal due to their drier conditions and favorable growing environment These regions provide the necessary climate for successful wheat production, ensuring a fruitful farming operation.
Table 9: Descriptive Statistics for Question 5, Layer evaluation
Layers Count Percent Correct Median Time
I conducted Pearson’s Chi-Square test to analyze the relationship between the number of layers and scores However, due to insufficient responses in each category, the test requirements were not met, leading me to accept the null hypothesis.
The second test investigated the correlation between the number of layers and response time, utilizing the Kruskal-Wallis test for analysis The findings revealed a Chi-Square value of 2.25 with 5 degrees of freedom and a p-value of 0.05, indicating an insignificant result Consequently, the null hypothesis was accepted.
I investigated the effect of a hillshade on student response time and score Table
10 presents the descriptive statistics for the mean score and response time for the two maps
Table 10: Descriptive Statistics for Question 5, Hillshade evaluation
Map Count Percent Correct Median Time
In my analysis, I employed Pearson's Chi-Square test to assess the correlation between hillshade presence and student scores The test produced a Chi-Square value of 0.526 with 1 degree of freedom and a p-value of 0.05, indicating an insignificant result Consequently, I accepted the null hypothesis.
The analysis revealed a minimal difference in mean response times, as shown in Table 10 Utilizing the Mann-Whitney U test, I found that the response times for maps with and without hillshades were statistically insignificant, with a Z score of -0.57, well below the critical threshold of 1.96 Consequently, I accepted the null hypothesis.
In Question 6, students utilized a dot density map to determine which county exhibited the highest level of a specific agricultural activity Each student received a randomly assigned question featuring between two and seven layers, focusing on various agricultural activities An example of this map, showcasing four layers, is illustrated in Figure 21.
Figure 21: Question 6 used dot density maps to symbolize quantitative information
The study aimed to assess how multiple layers influenced student response time and accuracy with varying symbolization types Only two tests were conducted, as the question did not include maps with hillshades Descriptive statistics for scores and response times are presented in Table 11.
Table 11: Descriptive Statistics for Question 6, Layer evaluation
Layers Count Percent Correct Median Time
Which county do you think grows the most onions?
In my initial investigation, I examined the correlation between the number of layers on the map and student scores The simplicity of the question may have led to a lack of expected incorrect responses, rendering the Pearson Chi-Square test ineffective Consequently, I utilized the results from the EXACT analysis, which revealed no significant findings, leading me to accept the null hypothesis.
I investigated the relationship between the number of layers and student response time Surprisingly, there was a decrease in the mean response times as the number of layers increased (Figure 22)
Figure 22: Question 6 response times per map There appears to be a decrease in response time as the number of layers increases
In my analysis, I utilized the Kruskal-Wallis test to compare response times to a question against the number of layers on the map The critical F value for degrees of freedom (df) = 4 at a significance level of p = 0.05 was determined to be 9.49, indicating that the results were not statistically significant Consequently, the variability observed between the groups could be attributed to chance, leading me to accept the null hypothesis.
Question 7 tasked students with analyzing a multi-layered dot density map to determine which agricultural activity was least dependent on precipitation The maps provided varied, with some students receiving a version that included a hillshade effect while others did not This inquiry aimed to assess the impact of hillshade on the interpretation and utilization of multi-layered thematic maps.
Due to a malfunction in one of the maps during initial tests in McMinnville, only eighty responses were collected, prompting the exclusion of those results After resolving the issue, I gathered fifty responses from students using the map without hillshade and thirty from those using the map with hillshade The descriptive statistics for response time and scores are presented in Table 12.
Table 12: Descriptive Statistics for Question 7, Hillshade evaluation
Map Count Percent Correct Median Time
Which agricultural activity do you think can happen regardless of how much rain there is?
Figure 23: Question 7 compared the effect of hillshade on the use of a multi-layered dot-density map.
I evaluated the relationship between presence of a hillshade and student response time The response times for the map with a hillshade and without a hillshade are nearly identical (Figure 24)
Figure 24: Response times for map with and without a hillshade for Question 7
In my analysis, I employed the Mann-Whitney U test to explore the relationship between response times for maps featuring hillshade and those without The test yielded a Z score of -0.33, indicating an insignificant result at p = 0.05, leading me to accept the null hypothesis.
I used a Chi Square test to evaluate any relationship between the presence of a hillshade and the score The Chi-Square value was 0.697, and the critical F-score for df
=1 and p = 0.05 is 3.84 Because the results were insignificant, I accepted the null hypothesis
Question 8 stood out from earlier questions by requiring students to analyze two maps and select the one they preferred for addressing a query related to precipitation and counties (Figure 25).
Figure 25: Question 8 presented students with two maps to choose from One with only the necessary layers to answer the question, and one with an extraneous hillshade
Look at the 2 maps below If you were going to pick one of these maps to use when answering questions about precipitation and counties, which one would you pick?
The study aimed to determine whether students preferred a map featuring hillshade, despite its irrelevance to the task at hand The maps were presented at a smaller size than earlier versions, as no specific tasks were assigned Ultimately, the majority of students opted for the map without hillshade, as illustrated in Figure 26.
Figure 26: Map preference among students in Question 8
The inquiry aimed to assess whether students preferred maps featuring hillshade, as these designs often appear more realistic and resemble commonly used reference maps Additionally, students were requested to articulate their reasons for selecting a particular map Their responses were subsequently categorized for analysis.
• Assumed it was the right answer
• Comment not applicable (e g “Because my cousin Ryan sucks.”)
In a survey of students who selected the hillshade map, 49% expressed a visual preference, while 26% found it more legible Additionally, 11% believed it was the correct choice, and 14% of the comments were deemed not applicable.
Figure 27: Reason given for picking the hillshade in Question 8
A significant majority of students, 75% (42 individuals), preferred the map without hillshade for its legibility, while only 7% (four students) believed it was the correct choice, and a mere 2 students attributed their preference to visual appeal.
Reason for picking map with hillshade
Figure 28: Reason student gave for picking map without hillshade in Question 8
Overall, 56% of students listed legibility for the reason they picked one of the maps over the other; 21% listed a visual preference.
In Question 9, students were tasked with selecting a map to analyze onion production across different counties, as illustrated in Figure 29 They had the option to choose between two maps: one displaying only the counties and onion data, and another that included additional information on precipitation and a hillshade layer.
Reason given for picking map without hillshade
Figure 29: Question 9 asks students to select a map to answer questions about growing onions and counties
When choosing a map to answer questions about counties and onion cultivation, it's essential to consider the clarity and relevance of the information presented Compare the two maps below and select the one that best illustrates the geographical distribution of onion-growing regions and their corresponding counties Your choice should enhance understanding of onion production areas and facilitate informed discussions about agricultural practices in those regions.
The two-layer map contained all the essential information needed to address the question, as illustrated in Figure 30 While most students preferred the map without hillshade, more than 40% opted for the version that included additional, unnecessary details.
Figure 30: Map Choices for Question 9
I also asked students a follow-up question about their choice I evaluated the responses and assigned the responses to the following categories:
Most students preferred the map without hillshade due to its enhanced legibility, while those who chose the hillshaded version expressed a preference for its visual appeal.
Figure 31: Reason given for choosing map without the hillshade
Figure 32: Reason given for choosing map with hillshade present
I did a Pearson’s Chi-Square test to compare map selection to stated reason and for choice and found a significant relationship between map choice and reason for choosing the map
Reasons for choosing map without hillshade
Reasons for choosing map with hillshade
Analysis and Discussion
The findings reveal a notable correlation between the number of layers in web-based layer maps and students' response times to questions This chapter analyzes the outcomes of each question, shedding light on what these results reveal regarding student interaction with the layer maps Additionally, some results prompted further inquiries, while others highlighted design issues in specific questions.
In Question 1, students were tasked with identifying the region that received the most precipitation using one of three different maps, each featuring a varying number of layers The findings indicate that the number of layers on a map significantly influences student response times, as shown in Table 13 The Kruskal-Wallis test results confirm that there are notable differences in response times among users of maps with different layer counts.
Table 13: Summary of Findings for Question 1
Compared Test used Test result
Layers & Score SPSS Exact 313 05 No
Layers & Time Kruskal-Wallis χ2 = 26.50 df=2, 5.99 Yes
This supports my research hypothesis that the effect of layers on response time, but there was no significant finding regarding the effect of layers upon student scores
The average score for the initial question was 0.89, reflecting a strong success rate among students Notably, there was no significant difference in the mean scores of students utilizing maps with varying layers—two, three, or five layers.
Evidence suggests that hillshade presence impacts response time, supporting one of my research hypotheses However, the hillshade test results may be influenced by the same factors as the layers test The map with five layers, which included a hillshade, showed significantly greater response times in the layers test This raises the possibility that the hillshade's effect on response time could be linked to the number of layers present, but insufficient evidence exists to definitively support this claim.
The absence of a strong correlation between hillshade presence and response times suggests that the number of layers was likely the primary factor influencing the significant findings in the hillshade and response time analysis.
There was no significant difference in mean scores for students who received a map with a hillshade and those who did not
Due to the straightforward nature of the question, there were insufficient incorrect responses for conducting the Pearson Chi-Square test on score-based comparisons Instead, the SPSS Exact test was utilized, which confirmed that the null hypothesis should be accepted.
Question 1 was the only multi-layer map test that indicated a positive relationship between the number of layers and the response time That is, as the number
As the number of layers increased, the response time also rose, but the results for subsequent questions did not follow the same trend This discrepancy may be attributed to the first question being the students' initial experience with the map, allowing them to familiarize themselves with it While later questions varied, this initial interaction likely influenced their performance However, there is no data available to confirm this hypothesis.
In Question 2, students were tasked with a similar activity as in Question 1, but with maps featuring varying layers: one map had four layers, two maps contained six layers, and one map included seven layers The findings presented in Table 14 reveal a significant correlation between the number of layers on a map and the response times of the students, supporting the research hypothesis that a notable relationship exists between the quantity of layers and the time taken to respond.
Table 14: Summary of Findings for Question 2
Factors Compared Test used Test result
Layers & Time (6 layer results combined) Kruskal-Wallis χ2 = 8.65 df=2, 5.99 Yes
Layers & Time (6 layer results separated) Kruskal-Wallis χ2 = 9.15 df=3, 7.82 Yes
The analysis of Question 2 reveals no significant relationship between the number of layers and scores, nor between the presence of hillshade and scores or response times Two maps with six layers were tested—one featuring hillshade and the other without—to assess any variations in scores or response times while keeping the layer count constant Both maps achieved identical mean scores of 0.94, and although the mean response times varied, the difference was not statistically significant, with a Z score of 0.76.
Unlike Question 1, there was not a positive relationship between the number of layers and response time (Figure 33), and the map with the fewest layers had the highest median response time
Figure 33: Response times for each map in Question 2 The map labeled 6C did not include a hillshade as one of the layers.
Upon reviewing the map with four layers, I found no clear reason for the higher mean response time compared to the seven-layer map The primary distinction between the two was the number of layers; the seven-layer map included the same four layers plus additional visual elements This discrepancy may indicate that certain layers can enhance the map reader's understanding by providing contextual information Specifically, the seven-layer map featured rivers, highways, and county borders, which likely offered valuable insights that aided students in answering the questions effectively.
Questions 3 and 4 required students to identify a specific county seat on the map and determine the precipitation level from four multiple-choice options Each question provided a brief description of the location, specifically noting the southwest corner of the state, to minimize the influence of students' prior knowledge on their response times when locating the county seats.
The findings reveal a significant correlation between the number of layers on the map and student scores, while no significant link was found between the number of layers and students' response times This outcome was unexpected, contrasting with the results from Questions 1 and 2 Additionally, the Chi-Square analysis indicated that the relationship between the number of layers and response time was just above the threshold for significance, as shown in Table 15.
Table 15: Summary of Findings for Questions 3 & 4
Compared Test used Test result
Layers & Time Kruskal-Wallis χ2 = 10.68 df=5, 11.07 No
The map with the fewest layers surprisingly yielded the lowest median score, particularly for the question related to McMinnville, Oregon, where 46% of the sample originated Despite including only the essential layers to answer the question, this two-layer map did not result in significantly higher mean scores for McMinnville students, as 12 out of 24 participants from this area performed similarly to their peers This counterintuitive outcome remains unexplained based on the collected data.
The analysis revealed no significant correlation between hillshade presence and student scores or response times in Questions 3 and 4, indicating that the research hypotheses related to hillshades are not supported.
My experience with the question randomization feature in Blackboard may benefit future research The software’s capability to randomize the delivery of maps in Questions 3 and 4 was instrumental in assigning two different maps effectively.
Concluding Remarks
This research investigated the impact of multiple layers on web maps on middle school students' question-answering abilities Initially, I hypothesized that an increase in thematic layers would lead to a decrease in the accuracy of student responses However, the findings did not support this hypothesis, leading to its rejection While questions 3 and 4 hinted at a potential significant relationship between the number of layers and student scores, the data did not demonstrate a decline in scores with an increase in layers; notably, the lowest scoring map for those questions actually had the fewest layers.
The findings indicate that while thematic maps should not be overloaded with layers, adhering to standard cartographic practices, incorporating additional thematic layers does not hinder students' ability to utilize the map for answering questions This suggests that educators should consider integrating multi-layer web maps to enhance their course curriculum.
The second hypothesis proposed a significant correlation between the number of layers on a map and student response time to questions Findings from at least two tests supported this hypothesis, although results varied across different questions Notably, there was no positive correlation; students utilizing maps with more layers did not exhibit a significantly increased response time.
The second hypothesis and its results were not meant to judge the significance of time in the use of educational maps While efficient map use can be beneficial in certain contexts, it does not reflect how students actually learn Although some response time differences were statistically significant, they lacked practical significance, as a few seconds may not impact classroom application Response time was utilized as a measure of map complexity rather than an indicator of student learning.
Research suggests that the number of layers on a thematic map significantly influences student response time This finding highlights the need for further investigation into how different types of layers impact map utilization.
The study did not assess the impact of various layer types and symbolization methods on response time Certain layers, such as roads, may enhance the user's ability to extract information, while others could hinder it Additionally, combinations of symbology, particularly maps featuring multiple layers with point symbols, might influence response times differently compared to maps that utilize diverse symbology types.
Bausmith and Leinhardt (1998) found that 7th grade students gained a deeper understanding of geographic phenomena through maps featuring multiple layers Feedback from students in Questions 8 and 9 revealed that they successfully identified connections between terrain and precipitation, highlighting their ability to recognize similarities between these two geographic elements.
The study aimed to investigate whether the presence of an extraneous hillshade affects students' use of web maps, hypothesizing that students would prefer maps with hillshades However, results from Questions 8 and 9 revealed that the majority of students (60% and 57%, respectively) chose maps without hillshades when they were not necessary, indicating a clear preference for simplicity Despite this, around 40% opted for maps with irrelevant hillshade layers, with some citing a preference for the more realistic appearance of these maps This finding raises concerns, as previous research has shown that hillshades can hinder map readers from accurately extracting non-landform information Consequently, the study explored the impact of hillshade presence on student accuracy and response time in subsequent hypotheses.
The fourth hypothesis suggested that students would respond more accurately to questions using maps without hillshade; however, the findings showed no significant difference in scores between the two types of maps This outcome may be linked to the nature of the initial questions, which focused on evaluating precipitation on a map Students might have identified a strong relationship between precipitation and the terrain depicted by the hillshade layer, potentially indicating that the hillshade enhanced their understanding of precipitation patterns.
The hillshade may have effectively served as a base layer in the mapping exercise, allowing students to focus on the relevant thematic layers without distraction By functioning as a background, the hillshade likely did not hinder the students' ability to analyze the primary information required for their questions This suggests that students could successfully disregard the hillshade and concentrate on the more pertinent data presented in the other layers.
If the terrain information had appeared unrealistic, such as through the use of neon colors in the hillshade, would students have been less inclined to choose that map? Additionally, could the bright colors have diminished their ability to ignore the hillshade's impact on their perception of the terrain?
The analysis excluded Question 6 due to design issues and its impact on response time, yet it offers insights into hillshade effects Notably, Question 6 utilized dot density maps without incorporating terrain or precipitation data, revealing no significant score differences between the maps with and without hillshade.
The fifth hypothesis proposed that students would respond faster to questions using maps without hillshading compared to those with it However, the findings did not support this claim, as there was no significant difference in response times between the two types of maps In fact, the average response times for both map types were almost the same, which was unexpected.
I cannot determine from my results whether students grew accustomed to the hillshade
During the test, it was observed that students either managed to ignore the visual distractions created by the hillshade layer or found it did not significantly complicate the map's usability Notably, around 40% of students expressed a preference for maps featuring hillshades, indicating potential for further exploration in this area.
After rejecting my fourth and fifth hypotheses, I find myself less concerned about the implications of my third hypothesis If hillshade does not adversely affect student scores or response times, and students can connect spatial phenomena such as precipitation with terrain, incorporating a hillshade layer in web maps could enhance student learning Therefore, I recommend that those creating web maps for middle school students, or educators utilizing tools like Google Earth, should include a hillshade layer to effectively represent terrain.