lllllllllllllllllllllllllllllll I 3 1293 00997 1460 This is to certify that the thesis entitled THEMATIC MAP LEGEND DESIGN FEATURES AND AUTOMATED CARTOGRAPHY ‘ .e**,med by Robert Michael Conley has been accepted towards fulfillment of the requirements for M. A. degree in Geography Major professor Date Mafi’//JI /97y 0.7639 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY M'Chlgan State University gm , SLE’JEL lag “i in Lew :4 ii? hula letri 1L. lll \ .u- r.-. PLACENREI‘URN my L n yowncord. TO AVOID FINES mum on or “on date duo. DATE DUE DATE DUE DATE DUE (an: (if) ‘1 I MSU I. An Minn-five Action/Emu Opportunity Intuition m THEMATIC MAP LEGEND DESIGN FEATURES AND AUTOMATED CARTOGRAPHY By Robert Michael Conley A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Geography 1 994 ABSTRACT THEMATIC MAP LEGEND DESIGN FEATURES AND AUTOMATED CARTOGRAPHY By Robert Michael Conley A thematic map legend design is crucial to the understanding of an intended map message. This study investigates the importance of various thematic map legend design features and then looks at how well computer mapping software packages manipulate those features when designing a legend. Fifteen legend design features were identified. A questionnaire was distributed to forty professional cartographers and geographers asking their opinions about relative importance of the fifteen features. Twelve computer mapping software packages were evaluated on their ability to design a thematic map legend by manipulating the same fifteen features. This study resulted in the ranking of the features according to their judged importance and the twelve computer mapping software packages according to their legend design performances. I dedicate this thesis to my loving family who gave me their full support and encouragement. iii ACKNOWLEDGMENTS I would like to acknowledge Dr. Richard E. Groop for assisting me with my thesis topic and for lending support when I needed it the most. Dr. Groop gave up several hours of his personal time for meetings and was always open to my ideas. I would like to thank Dr. Judy M. Olson for her thoughtful edits and input on my survey booklet design, and Bill Enslin at the Center for Remote Sensing for allowing me to use the CMAP GIS software and hardware. Finally, I would like to thank Nancy Yattaw and Jim Wilfong for their assistance in testing the graphics design for my survey booklet. iv TABLE OF CONTENTS LIST OF TABLES .................................................................................................. vu LIST OF FIGURES ................................................................................................ viii CHAPTER I. THEMATIC MAP LEGEND DESIGN ISSUES ............................................... 1 Legend Design Variables ................................................................................. 1 Legend Design Research ................................................................................. 5 Computer Mapping and Map Legends ............................................................ 9 Study Objectives ............................................................................................. 10 II. EVALUATING THEMATIC MAP LEGEND FEATURES ........................... ll Establishing the Importance of Legend Design Features ............................... l 1 Computer Mapping Software Package Evaluation ......................................... 15 Performance Rating Scale Development ........................................................ 29 III. ANALYSIS AND RESULTS .......................................................................... 31 Feature Importance Survey ............................................................................ 31 Results of the Legend Design Feature Importance Ranking .......................... 34 "Neither" Responses ........................................................................................ 36 Mapping Software Performance Ranking ...................................................... 36 IV. LEGEND DESIGN CONCLUSIONS AND RECOMMENDATIONS 41 Legend Design Features ............................................................................... 41 Future Legend Design Research ................................................................... 42 Computer Mapping Software ....................................................................... 42 Future Computer Mapping Software Research ............................................ 43 APPENDIX A Booklet Response Data ................................................................. 44 LIST OF REFERENCES ........................................................................................ 52 vi LIST OF TABLES Table l. Categorized Legend Design Features ..................................................... 1 1 Table 2. Legend Design Feature Pairings ............................................................. 12 Table 3. Coding Scheme for Questionnaire Booklet ............................................ 14 Table 4. Thematic Mapping Software .................................................................. 16 Table 5. Software Evaluation Rating Scale .......................................................... 29 Table 6. Legend Layout and Text Frequency Response Data ............................. 32 Table 7. ANOVA Using Legend Design Importance Response Data ................. 34 Table 8. Legend Design Features Ranked by Mean Responses .......................... 35 Table 9. "Neither" Responses ............................................................................... 36 Table 10. Software Evaluation Scores ................................................................... 37 Table 11. Software Performance Scores ................................................................ 38 Table 12. Software Ranking by Performance Evaluation Scores .......................... 39 Table 13. Software Evaluation Difficulties ........................................................... 39 vii LIST OF FIGURES Figure 1 Legend box arrangements ..................................................................... 2 Figure 2. Graduated circle arrangements .............................................................. 3 Figure 3. "Natural" isoline legend ......................................................................... 4 Figure 4. "Natural" dot legend .............................................................................. 4 Figure 5. Thematic map legend types .................................................................. 13 Figure 6. Base maps ............................................................................................. 14 Figure 7. A dot map produced with MapViewer ................................................. 17 Figure 8. PC Globe choropleth map .................................................................... 18 Figure 9. Atlas*GIS feature map ......................................................................... 19 Figure 10. CMAP shaded areal patterns ................................................................ 20 Figure 11. IDRISI areal shading patterns .............................................................. 21 Figure 12. MapInfo choropleth map ...................................................................... 22 Figure 13. Microstation PC feature map legend .................................................... 23 Figure 14. AutoCad legend .................................................................................... 24 Figure 15. FreeHand thematic map ....................................................................... 25 Figure 16. CorelDraw graduated symbol map ....................................................... 26 Figure 17. Harvard Graphics flow map ................................................................. 27 Figure 18. PowerPoint flow map ........................................................................... 28 Figure 19. Response means for legend design features ......................................... 33 viii CHAPTER I THEMATIC MAP LEGEND DESIGN ISSUES Over the past 30 years, cartographers have actively conducted research on how to improve the quality of thematic maps or the techniques used to design them. An exhaus- tive amount of research has been conducted on an array of map symbol types, symbol positions on a map, and symbol perception by map readers. The goal of this type of research has been to understand better how people read thematic maps and whether the intended map message is received by the map reader. Thematic map legend designs must also be crucial to understanding map messages since legends explain map symbols. A thematic map is unlike a general purpose reference map, which shows graphically, a real world representation of the individual features in a given area and may include terrain, vegetation type, buildings, roads and so on. If a legend is used on a general purpose map, it usually indicates a poorly designed and constructed map (Shelton, 1985). A thematic map, however, needs a legend to explain the different symbols because they are generally far more abstract. It is an unwritten but cardinal rule of the cartographer that any symbol that is not self explanatory must be identified in a legend. Furthermore, any symbol in the legend should appear exactly as it does on the map, drawn in precisely the same size, manner and complexity (Robinson, 1984). Legend Design Variables Traditionally, the size, shape, and placement of a map legend has depended on the entire map layout design based on some general and vaguely-defined principles of visual balance, size relationship, and aesthetic quality (Dent, 1972). Beyond these principles, the cartographer has a large number of variables and design decisions to deal with when l creating a thematic map legend. Some variables, such as location on the page and size, are general to all maps and others, such as the arrangement of features, are specific to particular types of thematic representations. The section below describes some of the more important legend variables. An important design consideration for choropleth legends is the arrangement of the boxes and their associated values. The typical choropleth legend box has a 3:5 ratio and is stacked vertically with the value description to the right of the box (Figure 1A). This arrangement is the best solution for placing labels with each box (Cuff, 1982). A hori- zontal arrangement, however, is sometimes necessary because of space restrictions, map configurations, or simply as a preferred arrangement by the designer. Three alternative horizontal designs illustrate how labeling can be confusing (Figure 18), how one can use extravagant amounts of space (Figure 1C), and how labels above or below can result in a more compact layout (Figure 1D). [:0549 (3 S044 2549 50-74 75-100 | |50-74 DE] 1:] [:1 [:1 75-100 024 25 49 5074 75100 Figure 1. Legend box arrangements. Another consideration is the location of larger and smaller values: should the largest value be at the top or bottom with a vertically stacked legend design, and on the left or right with a horizontally arranged legend design? Some individuals may associate the largest value and darkest color as being "heavier" and prefer to see it at the bottom. Other individuals may see the largest value and darkest color as "higher" and prefer to see it as the top legend box. There may also be conflicting preferences with respect to a horizontal arrangement. Likewise there may be preferences associated with alternative diagonal arrangements or "natura " legend designs where either a part of the actual map is dupli- cated or a representative arrangement is drawn up with the appropriate labels and title attached. Questions to reveal similar preferences can be asked of a graduated symbol legend: should the symbols be vertically or horizontally arranged and in what order? Both ar- rangements have traditionally been acceptable on thematic maps, with the horizontal (Figure 2A) more common than vertically arranged symbols (Figures ZB). When legend space on a map is restricted, or as a matter of personal preference, a nested circle legend can be used as a third design (Figure 2C). The decision is affected by the map size, configuration, legend placement space, preferences of the cartographer, and results of research that suggest that nested symbols are associated with greater difficulty in value 00. O 9 estimation. O A000 @ 13 Figure 2. Graduated circle arrangements. 0 o O O O Isoline legends are traditionally a very simple verbal statement of the isoline interval: "Each contour line equals 10 feet". However, some designers have recently advocated the inclusion of a "natural" legend (Figure 3) that more graphically describes the line and (sometimes included) shading symbols (Delucia and Hiller, 1982). The design of these legends offer many choices since there are few traditional rules governing them. 1250 500 200, Figure 3. "Natural" isoline legend. Dot map legends are very similar to those of isoline maps. Traditionally, a point symbol with a verbal description of it‘s value required few design decisions. But with the advent of "natural" dot legends and the use of variable scaling for eliciting more accurate judgements of values (Olson, 1977 ), legends were used to show various densities (Figure 4) and new considerations and variables are added. one dot = 50-225 people rural population persons/sq. mile Figure 4. "Natural" dot legend. Other kinds of thematic representations such as flow maps (graduated line symbols), cartograms, three-dimensional illustrations, or nominal data maps have equally diverse legend design considerations. As noted earlier, some legend variables apply to all kinds of maps. These include the legend size relative to the map and page size, location on the page relative to other map elements, wording of headings and other explanatory text, font size, font style, and the internal arrangement of differing legend elements. Taken together, legend design variables for specific map types and those that apply generally to all maps present the cartographer with a myriad of design decisions. Un- doubtedly, some decisions are made carefully and some are made without much thought; most are decided according to convention and subjective taste; and certainly, few are made on the basis of objective research results. Legend Design Research Surprisingly little cartographic research has focused on thematic map legends and few "rules" have evolved concerning their design. In W, Robinson (1980) recognized the importance of legend design: "the legend is a map component and is capable of being varied according to the principles of contrast. Assigning visual impor- tance is a problem that continuously faces cartographic technique. Components such as the legend boxes are not always of equal importance to a particular map or among similar types of maps. The legends of many maps are of little importance particularly when the symbols used are well known or self explanatory, however, in other cases, the legend box may hold the key that makes the map intelligible." However, WWII! (Robinson, 197 8), after 30 years as the leading textbook on cartographic practice in the US. and in its fourth edition, devotes only two paragraphs out of 448 pages to legend design issues. A book written by Cuff and Mattson (1982), MW Emdnclinn, addresses legends in some detail. They discuss the arrangement of legend symbols, especially graduated circles; choosing a legend heading; and they describe the positioning and size of legends on a map. While the book offers a more complete and thorough treatment of the thematic map legend, they base their design recommendations on convention and tradition, not experimental research conclusions. Over the past four decades, extensive and thorough research has been conducted on map communication, design, production, and use, and in particular, the perception quali- ties of thematic symbols. From these studies, many principles of map design, however tentative, have emerged. Few of these studies or resulting principles have dealt with legend design. And considering the importance of thematic map legends to their under- standing, this seems unusual. The majority of the research literature that directly exam- ines thematic map legends can be classified into four major categories: (1) natural legend designs, (2) bivariate classed and unclassed choropleth map legends, (3) color organization on two-variable maps, and (4) specific symbol perception studies. Delucia and Hiller (1982), conducted research on the potential of natural legend designs for a map as a more efficient means of communicating information. They re- marked “the legend of a thematic map is crucial to the map interpretation processes because the reader depends upon it to decode and comprehend the map symbols used. The symbol dimensions and patterns in the legend should be as identical as possible with those used on the map to avoid confusion and reduce inefficiency in the information communication process” (Delucia and Hiller, 1982, p. 46). Their findings indicated that a natural legend format (with layered tint maps) facilitated the reader’s ability to perform certain map reading tasks more efficiently and that overall map reading accuracy was measurably improved when a natural legend design was used. Fontanella (1989) studied the use of natural legends on battlefield maps. He noted that a natural legend depicts map symbols in a context more closely resembling actual situations within the body of a map, whereas a conventional legend catalogues them. He found this to be particularly appli- cable to battlefield symbols that are complex and highly abstract requiring a detailed legend and sophistication of the user. Neither of these studies tested design variables in the construction of natural legends, only the desirability of such legends over the conven- tional. Bivariate choropleth maps display, in overlay fashion, two distinct geographic vari- ables using enumeration unit data. These maps are both visually and intellectually complex and the assumption is that users must rely heavily on legends to correctly inter- pret them. Aspass and Lavin (1989) viewed unclassed bivariate choropleth legends as having the potential to communicate both data-specific and thematic information and that design of these legends may have an effect on reader's ability to acquire this information. After testing four legend designs using unclassed line symbols, they found that the presence or absence of a legend had little effect on map readers' abilities to identify regional trends. Brassel and Utano (1979) combined traditional discrete legend boxes (each designating a particular value) and Tobler’s (1971) continuous grey-scale legend which portrays the continuous character of unclassed choropleth symbols. They found that map readers were able to make numerical estimates from any map symbol despite the minimal information presented in the legend; however, “estimation accuracy has not yet been examined through empirical research” (Brassel and Utano, 1979, p. 41). Carstensen (1986) also examined bivariate choropleth map legend designs by using various methods of scaling the axes of the legends using continuous shading to represent two variables. The overall question posed by Carstensen concerned the relationship of the diagonals of the legend and the statistical trends in the data. After examining a sample of nine maps, he concluded that the axis scaling strategy had a distinct quantifiable impact on the effectiveness of the map. A number of other legend designs have been suggested for bivariate choropleth maps but their ability to assist map users in actual map use have not been tested. Lavin and Archer (1984) suggested using two separate legends, one for each variable and Smith (1977) designed a matrix legend that shows the bivariate relationship among three data classes for each variable using a combination of gray tone gradations and patterns. A third area of study involved the use of color on two-variable maps. The US. Bu- reau of Census generated and published several two-variable choropleth maps in the early 1970's. A paper by Olson (1975) detailed some considerations in the development of these color schemes. Olson suggested that: (1) colors must be distinguishable in the legend and map, (2) the color on a map must be matchable with the corresponding legend boxes, (3) the legend boxes must be distinguishable from each other and the color should progress smoothly and be visually coherent but separable, (4) a 4 X 4 legend might be visually subdivided into smaller categories (2 X 2), and (5) since readers associate darker colors with higher values, tones should progress from lighter to darker where the lightest tone should be at the lower left and the darkest the upper right. Olson (1981) conducted four experiments with human subjects to examine the efficacy of the census-style color two-variable legend. She concluded that: (1) subjects could not spontaneously order the colors into the legend arrangement used on the maps but they could recognize order in the arrangement; and (2) the legends on the maps were neces- sary to convey the relative values represented by the symbols. The fourth category of legend design research involves specific symbols. Dobson (1974) examined proportional circles and concluded that “the unwise selection of legend values may inhibit the transformation of information and render the map ineffective as a graphic display. In order to render the proportional circle map effective as an areal table, it is necessary to provide a series of circles in a legend that gives the reader an anchoring stimulus and assists him in estimating the magnitudes represented by the various sizes of circles on the map" (Dobson, 1984, p. 45). He defined the legend problem for a propor- tional circle map as how to provide the reader with a limited number of circles that can best define the values in the map body. The importance of thematic map legends was measured in an eye movement study by Antes, Chang, and Mullis (1985) where it was suggested that overall map design will influence the manner in which a map is initially viewed, which in turn, will influence the kind of information it communicates. They used a set of balanced and unbalanced gradu- ated circle maps where both contained a legend and other components that were logically and illogically placed. Their findings revealed that complex maps attracted more early attention to the legend than did simple maps. The longer the subjects viewed the maps, however, the less time they spent returning to the legends. The subjects also fixated longest on the subtitle (which revealed the thematic content) and second longest on the legend. It was concluded that map balance affects the distribution of attention only during early viewing, and that good map balance leads the eye to fixate readily on infor- mative map components like the legend. Computer Mapping and Map Legends Cartography today is rapidly and progressively becoming computerized. Computeriza- tion began with large mainframes but in recent years minicomputers and microcomputers have come to dominate automated mapping. Mapping software has reflected this trend moving from large special purpose mapping programs to small, general purpose pro- grams. Many automated mapping systems are parts of much larger systems such as statistical analysis or surveying systems. In addition, the growth of geographic informa- tion systems (GIS) has resulted in the development of integrated systems for the manage- ment and analysis of spatial data with computer mapping capabilities. Combinations of hardware and software now available for mapping often provide greater flexibility in compilation, design and symbolization than with traditional procedures. As these technologies have grown, more and more pe0ple are able to produce thematic maps, be they professional cartographers, GIS specialists, planners, or other public and private people working with spatial data. This means that the majority of the users of these products may have little or no cartographic training. It should be the responsibility of professional cartographers to monitor the quality of the maps produced and to attempt to influence the specifications of the hardware and software that produce them (Noronha 1987) As hardware and software systems for thematic map production have improved in recent years, design capabilities within these systems has also improved. Map makers l0 can now choose colors, select type size and fonts, arrange map elements on the page, select line weights, and make other types of design choices to enhance the map product. However, just as cartographic researchers have generally ignored thematic map legend design, mapping software developers often fail to include design flexibility in the creation of legends. While it is true that most computer-generated cartographic products do have legends, they are often very crude, hard to read and understand, and have poor visual esthetics, especially those from GIS software. The cartographer or GIS operator may meticulously design a map to his or her satisfaction and then, often as an afterthought, create a poor legend, sometimes because the software does not allow many design op- tions. Given the cartographic research record on thematic map legends and the proliferation of the software needed to produce legends for those maps, two questions arise: (1) what are the important legend design characteristics and (2) which mapping systems include legend design options and how well do they serve in the creation of good legends? Study Objectives This study focuses on specific thematic map legend design features and the ability of computer mapping software packages to manipulate these features when designing a legend. Included are the following two objectives: (1) To categorize essential thematic map legend design features and to determine the importance of these categories; and (2) using these results, to establish a set of criteria for judging legend design capabilities and to apply these criteria to computer mapping software packages. Results of the study will be a ranking of the importance of legend design features and of the goodness of software packages for designing legends. CHAPTER II EVALUATING THE IMPORTANCE OF THEMATIC MAP LEGEND FEATURES AND PERFORMANCE OF SOFTWARE LEGEND CAPABILITIES Establishing the Importance of Legend Design Features The first objective of this study was to categorize and determine the importance of the thematic map legend design features. After an extensive literature review of the text- books that discussed map legends and map design variables, a categorization scheme was developed that included two major categories: legend layout and legend text. These were subdivided into fifteen categories as shown in Table 1 below. Table 1 Categorized Legend Design Features Legend Layout Legend Text Legend size Clarity of all legend wording Legend shape Conciseness of all legend wording Legend location Type face Inclusion of a legend neatline Type style Visual proximity of legend features Type size Sequencing of legend features Text angle Arrangement of legend features Text location Text spacing ll Table 2 Legend Design Feature Pairings sequencing arrangement size shape location neatline proximity sequencing arrangement shape location neatline proximity sequencing arrangement size shape location neatline proximity location neatline proximity sequencing arrangement size shape location neatline neatline proximity sequencing arrangement size shape location proximity sequencing arrangement size shape spacing size arrangement TEXT F EAI URE PAIRINGS clarity conciseness type face type style type size angle location spacing conciseness type face type style type size angle location spacing clarity conciseness type face type style type size angle type face type style type size angle location spacmg clarity conciseness type face type style type size type style type size angle location spacmg clarity conciseness type face type style type size angle location spacing clarity conciseness type face angle location spacing clarity conciseness location spacing clarity 13 Next several thematic map legends were constructed for use as illustrations in the questionnaire booklet (Figure 5). These legends are the ones commonly used on the major types of thematic maps. They included: (1) natural legends for choropleth and isoline maps; (2) horizontally and vertically arranged conventional choropleth, (3) dot density; (4) graduated symbol; and (5) flow lines. COLLEGE FOOTBALL A'lTENDENCE (Odober 1993) .: b 100 000 101i 000 151,000 1.76 000 201 000 has 150 000 175.1110 200. 000 P H Graduated Symbol Peo e rs uare m'le . 1:5. 5:1 0‘: more People per square mile [:3 10110200 I ll 11 l[ 11 l :3 76 to 100 50 51 76 101 201 or to to to or E: 51 to 75 less 75 100 200 more C: 50 0' less Conventional Choropleth Conventional Choropleth People per square mile [:3 Urban 51975 1:: Forest [:3 Commercial morass D Rural [:1 Dunes 76b100 2; anarmure 10110210 ELEVATION Natural Choropleth (feet above sea level) ,0 "swi. ] 0 ONE DOT EQUALS 1O PEOPLE Dot Density Natural Isoline Figure 5. Examples of thematic map legend types. 14 These legends were chosen because they represented common legend types and because they provided considerable variety in appearance. They were alternatively attached to five different base maps (Figure 6) in a variety of positions and with different headings. The map-legend combinations were coded (Table 3), and the codes were Figure 6. Base maps. Table 3 Coding Scheme For Questionnaire Booklet BASE MAPS USED (code assignment): LEGEND TYPES USED (code assignment): . Natural (Bl) . Nominal [vertical] (32a) Nominal [horizontal] (BZb) . Dot Density (B3) . Isoline (B4) . Chompleth [vertical] (BSa) . Choropleth [horizontal] (BSb) Graduated Circle [vertical] (86a) Graduated Circle [horizontal] (86b) Flow Line (B7) . Shape Outline (B8) 1. United States (Al) 2. Europe (A2) 3. Australia (A3) 4. South America (A4) 5. Afiica (A5) ppgflthuN... hi _ _ 15 randomly assigned to all of the variables. Booklets were assembled with the constraint that the same base maps, legend types, or word sets did not appear on adjacent pages in the booklet. Two map-legend combinations were assigned to each booklet page such that a legend feature alternated between top and bottom (see sample booklet inside back cover). The random scheme reduced the possibility of learned responses and boredom. For the layout category, each feature was displayed in the booklet 6 times and for the text category each feature was displayed in the booklet 7 times. Thus, a person selecting "legend size" as the important feature four times in the booklet would be choosing the legend size feature four out of a possible 6 occurrences. The questionnaire booklet had 52 pages at a size of 8.25" X 5.5" (see sample booklet inside back cover). A page of instructions briefly explained the purpose of thematic map legends and how to assess feature importance. Each page had one feature pairing from the seven features of the layout category and eight features from the text category. At the bottom of each page, a place was provided to place an X or check mark for those repondants who could not decide between the two displayed features. Forty people were selected to fill out the questionnaire booklets. The people chosen were Michigan State University Geography Department faculty, staff, and graduate students, employees selected from the Center for Remote Sensing, and employees from the State of Michigan Land and Water Management’s MiRIS program. All were well qualified as map designers and in the application of cartography to computer mapping, remote sensing and GIS projects. Thus, participants were an "expert" set. Thirty-two response booklets were collected. The feature pairing responses were entered into re- sponse occurrence spreadsheets and resulting frequencies tallied. From these frequencies, the importance of each legend design feature was determined. Results are presented in the next chapter. Computer Mapping Software Package Evaluation To evaluate legend design capabilities of mapping routines, a survey was conducted of computer software for thematic mapping. From this survey, twelve packages were l6 selected for evaluation (Table 4). They were chosen on the basis of several qualities: ( l) availability and accessibility; (2) representation of different mapping tasks such as GIS applications, computer aided design, computer mapping, and presentation graphics; and (3) ease of use on a PC-type platform. In addition, the twelve programs represented a range of cost and sophistication. As such, these programs form a representative set of computer mapping systems most commonly used in business, public sectors, and aca- demic realms. Table 4 Thematic Mapping Software Software Version/Platform Price Manufacturer MapViewer 1.1/Windows 8 249.00 Golden Software, Inc. PC Globe 2.0/DOS 8 99.00 PC Globe, Inc. Atlas*GIS 1.1/DOS $ 2495.00 Strategic Mapping, Inc. CMAP 2.Beta/DOS $ 750.00 Michigan State University IDRISI 4.0/DOS 8 100.00/student Clark University Maplnfo 2.0. l/Windows 8 995.00 Maplnfo Corp. Microstation PC 4.0/DOS S 200.00/student Intergraph Corp. AutoCad I 1.0/DOS $ 495.00/student Autodesk, Inc. FreeHand 3.1/Windows $ 395.00 Aldus Corp. CorelDraw 3.0/Windows $ 395.00 Corel Systems Corp. Harvard Graphics 1.01/Windows S 125.00 Software Publishing Corp. PowerPoint 2.0/Windows 8 295.00 Microsoft Corp. Each software package was run on a 486-33MHz machine using Windows 3.1 and PC Tools on Top for Windows as the desktop. The images were generated by using Hijaak 2.0 for Windows and DOS to capture the entire monitor screen from each software package after a map and legend was made. This method allowed the software map 17 making environment to be displayed. The images were then saved in a Tiff file format and placed in Aldus Page Maker 5.0 for Windows. Following is a brief description of each program along with a sample thematic map and legend produced on each. MapViewer is a thematic mapping package that runs under Microsoft Windows. The software package comes with, sample base maps and data, and a 483-page manual. MapViewer is capable of producing several thematic maps including choropleth, gradu- ated symbol, and dot density (Figure 7). A default legend can be selected and then modified freely, or a custom legend can be created with various drawing tools. MapViewer is easy to learn and use and has text and line editing functions. Gallery floundary Set window Help BIRTHS 1 980- 1 986 Elle Edit [few Draw .I.l.1. .r.l.1.1.1.l.r. .1.l.r.I.r.l.1. I -' N . 3,3533: 1.1.].1.1.1.1.1.l.r.I.1.3.r. .1.l.r.i.1|1 35°" 6“ 33$? M. 5 t \ Figure 7. A dot map produced with MapViewer. l8 PC Globe is a DOS platform choropleth mapping software package (Figure 8) de- signed primarily for educational settings. PC Globe has a large data base and several base maps to choose from. This software uses only a default legend which cannot be edited. POPULATION 2888 (In thousands) 12.888-58.888 3.888-12.888 BBB-3.888 Above 888.888 288,888-888,888 58.888—288.888 Figure 8. PC Globe choropleth map. 19 Atlas*GIS is a vector GIS package that includes many complex mapping, editing, and data handling routines and comes with sample base maps and data sets. Most maps have two default legend spaces allocated (not shown in Figure 9) on the right side of the map area and these legends can be edited to an extent. Atlas*GIS offers a freehand drawing system which was used in the example. This software is mouse driven but has a steep learning curve and requires several steps to perform almost all functions. ATLASnkGlS Des atop (31110911313301: Enforn'letécn System F eafures E] States —- Interstates 1: Cities Figure 9. Atlas*GIS feature map. 20 CMAP, is a university-developed vector GIS software package that has extremely strong data conversion programs and utilities and modest mapping capabilities (Figure 10). CMAP uses a default legend; however, a "key" file must be created first. The key file contains the legend symbol descriptions and codes for the symbols and patterns to be used. After the key file is created, the map and legend can be displayed and a legend title created through a series of complex steps. CMAP also allows freehand functions. Ap- proximately one half of the commands and modules are mouse driven. HETLRND PIJIN‘IS Figure 10. CMAP shaded areal patterns. 21 IDRISI is a raster based GIS software package (Figure l 1). Also a university devel- oped product, it offers only a default legend with editing limited to the symbol descrip- tions. This is done through a text editor provided in the software or through an external text editor. IDRISI is mouse or keyboard driven and requires several complex steps to operate. Figure 11. IDRISI areal shading patterns. 22 Maplnfo is a GIS software package that has extensive mapping capabilities and runs under Microsoft Windows (Figure 12). This software includes a database, sample base maps, address matching, and point and boundary analysis. It offers several thematic mapping capabilities. It has a default legend (Figure 12A) with limited editing capabili- ties and some drawing tools for customizing legends (Figure 12B) and it is mouse driven. e table at Murrow ..... VIII-i 'I I‘- II.- .y _.'. '5.:.:.‘ . . . . . .-’-!-.,.- -. . v.- . .. w- , a". 1.0-:_:.I_: .. - 31-15-1216th 3- o' 'd'n‘ ;.:.-‘.:-;.r are!I 71: :5=:- 35:? .': .532. .4- £33.50; 'o'l:uVl :1 .-'='=:-. 55?“:- I-_.. 'u. Afifin‘uf’. ':" ' £31. .é. fit 13;; n‘ v!»- fill-KOD- .g. 1.1.1: «15112.4 asvns“,r°w1:-;-:-'5'-:»'-:-'3'-'-:::-:f: .;.;._._.-.-.;.:.;.-.‘.;. ;.;3:._._<'1:.:.;.;. ' ' (“15333-14'?’ - .-.-.-.-z.-.; "-. aux-z.- -.t. 2+. ‘ g "ml" oar-o . O O o ' 2...; . ff 5. isoot moor not 5000 15000 ioooo 1500 W5, Basketball New _ . . if??? ....... Figure 16. CorelDraw graduated symbol map. 27 Harvard Graphics is a presentation creation software package that has extensive drawing tools and capabilities (Figure 17). It comes with an extensive clip art gallery that includes a map section, and it has text and line editing capabilities. The program is mouse driven and is easy to learn and use. Harv ml [impluic 3 n I 533,3 3:353:33 Elie gait mew §llde Qhart Iext graphics window Help _ 7 _ . 3 mm Aims; Mlgratlon patterns (people pervear) Figure 17. Harvard Graphics flow map. 28 PowerPoint (Figure 18) is presentation creation sofiware similar to Harvard Graphics. It comes with clip art, is mouse driven, offers good text and line editing capabilities, and has extensive freehand tool functions. Microunlt l‘uwcrl’uinl H’Al UAlLJ'l’l Elle Edit Ilew Iext Qb|ect Arrange Slide window flelp Migration of people F 100,000 or more 4—— 0 to 99,999 Figure 18. PowerPoint flow map. 29 Performance Rating Scale Development To evaluate the legend design capabilities of the twelve software packages, 3 simple rating scale was developed to apply to the layout and text features identified above. The scale is outlined in Table 5. Table 5 Software Evaluation Rating Scale PERFORMANCE RATING SCALE: 0 - Cannot manipulate the features when designing a legend 1 - Difiicult to manipulate the features when designing a legend 2 - Easy to manipulate the features when designing a legend EASY AND VERSATILE: DIFFICULT AND INFLEXIBLE: -obvious functions and drawing tools -no obvious functions and drawing tools -few steps involved -several steps involved -several manipulations allowed -few manipulations allowed -default legend templates used -no default legend templates used -tool bars displayed -no tool bars displayed -button cursor or mouse driven -keyboard use only -freehand editing allowed -no freehand editing allowed -online help or tutorials -no/limited online help or tutorials -good reference materials -poor/limited reference materials An evaluation scale was applied to each of the twelve mapping packages. A variety of maps were constructed in each package and a subjective evaluation was conducted based on previous experience in producing these maps. A preexisting base map, usually Clip art, was used when possible. Each of the 15 features was manipulated in some fashion with the various drawing tools and available commands. A number was then 3O assigned from the rating scale for performance with that feature. Resulting ratings were used in conjunction with the design feature importance survey results to produce an overall rating index. The results of the software rating and the survey results are pre- sented in the next chapter. CHAPTER 3 ANALYSIS AND RESULTS Feature Importance Survey Questionnaire response data from the 32 returned questionnaire booklets was tabulated in the manner shown in Appendix A. Rows represent feature pairing combinations and each column is the chosen response to that particular combination. Each of the two categories (layout and text) had different potential response occurrences. In the layout category, every feature was paired a total of 6 times. In the text category, every feature was paired a total of 7 times. These responses were then tabulated as frequencies ( the number of times a respondant selected a feature) for each of the 32 booklets (Table 6). The "neither" response was ignored during the ranking process and was analyzed sepa- rately. To obtain an overall importance ranking for each legend design feature, mean re- sponse rates were calculated. Because the two categories had different possible total occurrences (layout-6; text-7), the features in each category were ranked within that category. Figure 19 shows the mean responses (excluding "neither" responses) for each feature in each of the two categories. An analysis of variance (AN OVA) was performed to test for a difference between the means. It was applied to each of the two categories and results are shown in Table 7. In each of the two categories, the critical F statistic from the ANOVA was less than the calculated F statistic, resulting in the rejection of the null hypothesis. In other words, the means in each category were statistically significantly different. 31 32 Table 6 Legend Layout and Text Response Data Frequency of Responses Legend Layout Design Features 7 Frequency of Responses Legend Text Design Features Response Frequency Response Frequency 4. , ' __ ‘ Carotene“ Text Lee-eon Type st. Legend Text Features 33 /#fi Mean Responses ‘ Legend Least-en A Legend Sue” Legend Layout Features (if! Mean Responses Sequencng Legend Shape Nuthe MA, Two SM. Text Specng Type Flee .2“ . Text Anw Figure 19. Response means for legend design features. 34 Table 7 ANOVA Using Legend Design Importance Response Data Anova: Single-F actor Summary Legend Lem Fearins Cour! Sum Mean Response Variance Legend Ste 32 92 2.88 2.5 Legend Shape 32 47 1.47 1.48 Legend Location 32 105 3.29 1.31 Inclusion ot a legend Neetline 32 22 0.69 1.25 Visual Prox‘mlty of Legend Features 32 130 4,06 1.87 Sequencing of Legend Features 32 91 2.84 3.1 Arrangement a! Legend Features 32 123 3,84 1.04 ANOVA Source of Variation SS d! MS F P-velue F or! Between Groups 291.71 6 48.62 27.11 4.960016-24 2.14 Within Groups 389.13 217 1.79 Total 680.84 223 Anova: Single-F actor Summary Legend Text Foams: Com! Sum Mean Response Variance Legend Text Angle 32 92 2.88 3.73 Legend Type Size 32 106 3.31 2.16 Legend Type Style 32 22 0.69 0.8 Legend Text Spacing 32 79 2.47 1.29 Legend Type Face 32 33 1.03 1 77 Legend Text Location 32 139 4. 34 1 2 Clarity of Legend Wording 32 185 5.78 1.53 Conciseness of Legend Wording 32 148 4.83 2.18 ANOVA Source of Variation 85 at MS F P-value F at Between Groups 692.56 7 98.94 54 2 097218-46 2.05 \Mth’n Groups 454.38 248 1.83 Total 1146 94 255 Results of the Legend Design Feature Importance Ranking Table 8 show the importance of the legend design features ranked by the mean re- sponses for both the legend layout and legend text categories. The higher number means that feature received "more important" responses in the feature pairings and the lower number means it received fewer responses by the 32 people. The similar ranking of the proximity and arrangement of legend features is logical in that they are both related to the placement of features in a legend. The lower ranking of legend shape probably indicates that the shape is a result of the influence that other features have on it. In other words, manipulating the other legend design features might determine the shape. Finally, the 35 legend neatline is ranked last, probably because it might not be viewed as a necessary legend design feature but rather as an artistic addition. Table 8 Legend Design Features Ranked by Mean Responses Legend Layout Mean Response 1. Visual Proximity of Legend Features 4.06 2. Arrangement of Legend Features 3.84 3. Legend Location 3.29 4. Legend Size 2.88 5. Sequencing of Legend Features 2.84 6. Legend Shape 1.47 7. Inclusion of a Legend Neatline 0.69 Legend Text Mean Response 1. Clarity of Legend Wording 5.78 2. Conciseness of Legend Wording 4.63 3. Text Location 4.34 4. Type Size 3.31 5. Text Angle 2.88 6. Text Spacing 2.47 7. Type Face 1.03 8. Type Style 0.69 With text features, clarity and conciseness are both ranked highly and are probably seen as being interrelated. The two features type face and type style, ranking last, were the two least important features and could be veiwed as a non-functional aspect of the legend design. 36 "Neither " Responses The "neither" responses were not used in the initial feature response analysis. How- ever, it is important to understand why these responses were given and whether the results can be used in some meaningful way. Almost all feature pairings had low "neither" response rates (5 out of 32). Only four pairings had higher "neither" response rates than this norm. These are shown in Table 9. Table 9 "Neither" Responses Feature Pairing Response Percentage clarity-conciseness of wording 15 text location-angle 7 type style-type face 6 text angle-spacing 6 These responses may be the result of two possible explanations: (1) Either respon- dents were not able to differentiate the features (i.e., the difference between type style and type face), (2) the features are so interrelated that respondents could not decide on which was most important, or (3) the features are equally important. Mapping Software Performance Ranking The legend design capabilities of each software package were subjectively evaluated using the performance rating described in Table 5. The results are presented in Table 10. There was considerable variation among software packages in the flexibility allowed for legend design. In general "graphics" and "presentation" software allowed the most Table 11 presents the combination of performance ratings and design feature impor- tance rankings. The ranking of each feature was multiplied by the performance rating to create a score for each software package in each category. These were then summed for the software to provide a composite rating. Table 12 lists the packages in order of com- posite rating. 37 design choices while "GIS" and "CAD" packages had more flexibility. Overall, the twelve software packages used here seemed to provide a full range of capabilities. Table 10 Software Evaluation Scores Ialeewer 1.1 PC Globe Adas'Gls 1.1 (MM? 2.Bete IDRISI 4.0 Thematic GIS ext ext Face Location 2 2 Maplnfo 2.0.1 Microstation 4.0 AutoCad 11.0 FreeHand 3.1 Corele 3.0 CAD CAO of Features of ext ext Face Location of Presentation LA Location of a Neatline of Performance Rating: of 0 - Cannot manipulate the features to design a legend 1 - Difficult to manipulate the teamres to design a legend 2 - Easy to manipulate the features to design a legend 38 Table 11 Software Performance Scores Loation Inclusion 01 a Location lnciusmn of a Neatline of F of F ext Location of TOTAL Face ext Location TOTAL The highest number is the best overall software package allowing the most flexibility when manipulating the individual legend design features and the lowest number offers the least. Harvard Graphics and Corel Draw received the highest scores, both at 128 points. FreeHand ranked next with 123 points and PowerPoint and Map Viewer both 39 scored 118 points. A summary of software legend design difficulties appears in Table 13. Table 12 Software Ranking by Performance Evaluation Scores Software Ranking By Performance Scores Harvard Graphics 128 Corel Draw 128 F reeHand 123 MapViewer 118 PowerPoint 118 Microstation l 09 AutoCad 94 Maplnfo 87 Atlas*GIS 84 CMAP 72 IDRISI 9 PC Globe 0 Table 13 Software Evaluation Difficulties Software Package Layout Difficulties Text Difficulties Harvard Graphics none none CorelDraw none none FreeHand none none MapViewer none cannot manipulate text angle PowerPoint none cannot manipulate text angle MicroStation steep learning curve manipulating type style AutoCad steep learning curve manipulating type style and face Maplnfo none cannot manipulate text angle Atlas‘GIS steep learning curve steep learning curve CMAP awkward manipulations/steps manipulating type style and face IDRISI no manipulations allowed limited text editing PC Globe no manipulations allowed no manipulations allowed Both of these tables can provide a usable guide (or a basis for recommendation of appropriate software) for the users of software who are concerned with thematic map legend design flexibility. 40 In general, software that uses the windows environment, especially those with graphics capabilities, ranked higher than those in the DOS environment with command driven menus and interfaces. The Graphics, CAD, and Presentation packages have the best overall performance. The two operating systems themselves, DOS and Windows, do not seem to be a direct influence on the ratings, however; as long as the package offers some sort of graphics capabilities it ranks more highly. Maplnfo is the exception; there is graphics capability, but it is limited in the types of manipulations allowed. Finally, locally develOped university software such as CMAP and IDRISI have low legend design capabilities. PC Globe, on the other hand, is a commercial package and has no legend design capabilities. CHAPTER 4 LEGEND DESIGN CONCLUSIONS AND RECOMMENDATIONS Good map design has long been a major research field in cartography. Most of the emphasis has been on individual map components. Since then, cartographic research has moved in the direction of visualization, understanding geographic patterns "the whole map picture" and other research arenas. During the same time period, cartography has been revolutionized by automation which has brought about a need for evaluating and investigating automated designs in terms of their cartographic quality. This thesis evaluates the legend designs produced by computers by identifying the design features important to professionals to aid the user in designing automated maps. Computer mapping software products are designed by programmers with little back- ground in cartography. Thus, their software designs for map making vary in flexibility in terms of a users ability to manipulate legend design features. One cannot overlook the fact, however, that even the most flexible software cannot produce a good map design unless users themselves have the knowledge or guidance to produce an appropriate map design. This thesis has produced a thematic mapping software selection list (based on flexibilty) and a ranking of 15 legend design features based on importance that can be used by the map maker. Legend Design Features This study has shown that some legend features are more important than others. The results of the questionnaire on legend design features has indicated the relative importance of these features and has revealed that some of these features are more impor- tant than others. Table 9 can be used as a guide when designing a thematic map legend 41 42 regardless of the type of legend being made. It shows that the visual proximity and the arrangement of legend design features are particularly crucial and should be included on all legend designs while the inclusion of a legend neatline can be considered less impor- tant (according to the experts questioned in this study) and the designer can choose whether or not to include one around a legend. Future Legend Design Research The input in this study has been provided by relative experts in cartography who completed a questionnaire. The results provide food for thought concerning further research. Work should be done by developing a series of testing instruments and to use these instruments to test map users on the importance of thematic map legend design features in hopes that maps can be read in a more efficient and meaningful manner. Research questions could center on the arrangement and the sequencing of legend design features (particularly with choropleth and graduated circle map legends); horizontal vs vertical legend arrangements; whether or not there is a preferred sequencing of the choropleth legend boxes and graduated circles; and whether map users prefer the larger value and darker colors to be on the top with vertically arranged choropleth legend boxes and graduated circles, and on the right with horizontally arranged boxes and graduated circles. The results of these studies combined with Table 9, could more definitively guide thematic map legend designers. Computer Mapping Software This study has shown that there is considerable variability in the ability of computer mapping software to handle thematic map legend design needs. In general, the CAD and freehand graphics software packages performed better than the GIS and presentation software. On another dimension, the common commercial software, as was expected, performed better than the packages that were University developed (Table 13). The difficulties in the different packages varied considerably (Table 14). 43 Future Computer Mapping Software Research The twelve software packages evaluated for this study were used on a PC computer in the DOS or Windows environment. The same or similar software packages should be evaluated in a UNIX and Macintosh environment as well. Some examples of software for the UNIX environment would be Intergraph's MGE/MGA, Grass, and Arc/Info. For the Macintosh environment, Map 11, Illustrator, and Map Grafix could be evaluated. Other software for the PC DOS environment that could be evaluated includes ERDAS, Surfer, and Arc/Info. A more stringent and standardized set of evaluation criteria could be developed and applied when evaluating software in any environment. This criteria could include the flexibility of the import and export functions offered by the software, output capabilities such as laser printing and pen plotting, and default legends (if any). Finally, a set of software purchase or use criteria could be developed as a result of platform and operating environment comparisons for thematic map legend design. APPENDIX A 44 APPENDIX A Booklet Response Data 45 46 47 48 49 50 51 LIST OF REFERENCES LIST OF REFERENCES Antes, James R., Chang, Kang-tsung, and Mullis, Chad. 1985. “The Visual Effect of Map Design: An Eye-Movement Analysis.” The American Cartographer, vol.12, no.2, pp. 143-155. Aspaas, Helen Ruth and Stephen J. Lavin. 1989. “Legend Designs for Unclassed Bivariate Choropleth Maps.” The American Cartographer, vol. 16, no. 4, pp. 257- 268. Carstensen, Laurence W. 1986. “Bivariate Choropleth Mapping: The Effects of Axis Scaling.” The American Cartographer, vol. 13, no. 1, pp. 27-42. Clark, W.A.V. and Hosking, PL. 1986. Statistical Methods for Geographers. John Wiley & Sons, Inc., New York, NY., 518. Clarke, Keith C., 1990. Analytical and Computer Cartography. Prentice Hall, New Jersey, 290. Cuff, David J. and Mark T. Mattson. 1982. Thematic Maps: Their Design and Production. Methuen & Co., New York, NY., 169. DeLucia, AA. and D.W. Hiller. 1982. “Natural Legend Design for Thematic Maps.” The Cartographic Journal, vol. 19, no. 1, pp. 46-52. Dent, Borden D. 1972. “Visual Organization and Thematic Map Communication.” The American Cartographer, vol. 2, no. 2, pp. 79-93. Dobson, Michael W. 1974. “Refining Legend Values for Proportional Circle Maps.” The Canadian Cartographer, vol. 11, no. 11, pp. 45-53. Fontanella, Joseph F. 1988. Assessment of the Eflectiveness of Battlefield Maps. East Lansing: M.S.U., (Masters Thesis). Gilmartin, Patricia P. 1978. “Evaluation of Thematic Maps Using the Semantic Differential Test.” The American Cartographer, vol. 5, no. 2, pp. 133-139. 52 53 International Paper Co. 1989. Pocket Pal: A Graphic Arts Production Handbook. International Paper Co., Memphis, Tn., 220. Keates, John. 1973. Cartographic Design and Production. Second Edition, John Wiley & Sons, Inc., New York, NY., 240. MacEachren, Alan M. 1987. “The Evolution of Computer Mapping and its Implications for Geography.” Journal of Geography, May/June, pp. 100—107. Marble, Duane F. 1987. “The Computer and Cartography.” The American Cartographer, vol. 14, no. 2, pp. 101-103. Michigan State University. 1993. The Graduate School Guide To The Preparation 0f Master's Theses And Doctoral Dissertations. East Lansing, Michigan, 32. Monmonier, Mark S. 1982. Computer-Assisted Cartography: Principles and Prospects. Prentice Hall, New Jersey, 214. Muehrcke, Phillip. C. 1972. “Thematic Cartography.” Association of American Geographers, (resource paper no. 19), 66. Muehrcke, Phillip C. 1990. “Cartography and Geographic Information Systems.” Cartography and Geographic Information Systems, vol. 17, no. 1, pp. 7-15. Noronha, Valerian T. 1987. “Choropleth Mapping in a Microcomputer Environment: A Critical Evaluation of Some Commercial Implementations.” The American Cartographer, vol. 14, no. 2, pp. 139-154. Olson, Judy M. 1975. “The Organization of Color on Two-Variable Maps.” Auto Carto II Proceedings, pp. 289-294. Olson, Judy M. 1977. “Rescaling Dot Maps For Pattern Enhancement.” The International Yearbook of Cartography, vol 17, pp. 125-137. Olson, Judy M. 1981. “Spectrally Encoded Two-Variable Maps.” Annals of the Association of American Geographers, vol. 71, no. 2, pp. 259-276. Paslawski, J. 1983. “Natural Legend Design for Thematic Maps.” The Cartographic Journal, vol. 20, no. 1, pp. 36-37. Petchenik, Barbara Bartz. 1974. “A Verbal Approach to Characterizing the Look of Maps.” The American Cartographer, vol. 1, no. 1, pp. 63-71. Robinson, Aurther H. 1980. The Look of Maps: An examination of Cartographic Design. The University of Wisconsin Press. Madison, Wisconsin, 105. 54 Robinson, Aurther H. and Randall Sale. 1978. Elements of Cartography. Fourth Edition, John Wiley & Sons, Inc., New York, NY., 270. Turabian, Kate L. 1987. A Manual for Writers of Term Papers, Theses, and Dissertations. Fourth Edition. The University of Chicago Press. Chicago, Illinois, 300. Shelton, Hal. 1985. Video Tape. Map Division, Library of Congress. Strunk, William and White, EB. 1979. The Elements of Style. Third Edition. MacMillan Publishing Co., Inc. New York, New York, 92. Werner, Robert and Young, James. 1991. “A Checklist to Evaluate Mapping Software” Journal of Geography, May/June, pp. 118-119. nrcnran smrs UNIV. Lramnrss llllllmlllllllllWNWINlllllllllllllllll”"Illllllll 31293009971460 e .rl. a n n .m t. S e m n .Wo S e D d n e at e L P a M .m t a. m e In T s. w? QEQMVMWNMW Y -M. Conley l. algaL‘ga_..‘.__fl.__,._:_z xterm by) f ’ l ‘n [W can“, ‘ ‘va’m ll Lauri. lBRARlES L MICHIGAN STATE UNNERSFW LANSING. MlCH 438%. 11.33 When constructing map legends, cartographers make many choices. Some are probably more important than others. i would like your opinion of the relative Importance of some of those decisions. LOOk at the pairs of items on the following pages and tell which choice in each pair is a more important one in constructing a good map legend. in other words, Which choice is more important to the functioning of the legend (and hence the map), Considered from another angle, with which choice would a BAD selection be most detrimental to the functioning of the legend (and the map). The term "functioning" can have different meanings. At the lowest level, a legend feature must promote und§§adtyrtnabili and speed pf compeerhneion. A map a in fr 3 ' legend should also r u i v i of ma re and have an H ar n if it is "functioning" well. YOU may find some pairs of choices to be of equal importance. l have provided SUCh an answer ("Neither; they are equally important"), but I hope you will use it _°”'y When you can see no reason for one of the choices to be considered more 'mportant. Please mark your answers with an X directly on the booklet pages. ........ J \ Legend Shape \ Neither; they are equally important \ Type Size Land Use/Cover E I: E l: E OR Urban Foresl Commercial Rural Dunes \ Conciseness of (3”) Legend Wording DEEDS] Urban For.“ Commercial Rural Dunn \ Neither; they are equally important Jrjr Which is more important? Land Use/Cover [:1 Urban Land Use/Cover l:]Urban Forest E Urban: \ . [2 Forest 2 Commercial Forest :I \ Text Locatlon [: Commercial Rural l: Commercial E :Rural :IDunes Rural: l: Dunes Land Use/Cover Dunes '2 Urban grbam Urban Forest ores . Fcnrest \Text Spacing Commercial Eggrlnercral C 0 m m e r C i a I Rural Dunes RlJral u n e s Dunes D \ Neither; they are equally important Which is more important? \ Legend Size OR A Arrangement of Legend Features V 'l ‘ I‘ - 4'; d . ; ONE DOT EOUALS 10 PEOPLE 0 ONE oor EQUALS 10 PEOPLE o Neither; they are equall im «ortant Which is more important? \ Sequencing of Legend Features Peopepersqzemle - 20$ ormore - 101 to 200 - 7510100 51 «075 OR ‘ I: Scorless \ Legend Size Peoplepersquarem'ie - 201 armors - 101lo200 - 7610 100 511075 ‘ C: SOorIess \ Neither; they are equally important I: 500rless 511075 - 76l0100 - 101 10200 - 201 orrnore ‘ Peoplepersquarerrile Which is more important? \ \ Conciseness of (all) ’ Legend Wording : ' ‘ SEEK/CW I ‘2 Urban Land Use E] ForeleInd Cover E] Commercial Land Use D Rural Lend Use 0R C] Dune': Land Cover ti \ Text Spacing LandUse/Cover Land Use/Cover- .\ Neither; they are equally important 1- \ Inclusion of a Legend Neatline OR \‘ Visual Proximity of Legend Features (how close the symbols, descriptions, and title are to each other) _. \ Text Location ONE DOTL EQUALS ' ' ONE DOT EQUALS E O 10 PEOP 1o PEOPLE 0 OR (Conciseness of (all) Legend Wording \ Neither; they are equally important Which is more important? People per squire mile 7g Clarity of (all) Legend Wording O R Pemle per square mile 2010! more 10‘ lom 1 \ Legend Location People per mun mrle 201 w mm Htfl‘lbm \ Neither; they are equally important Which is more important? Land Use/Cover / . \ Concrseness of (all) g 3”" oresl Legend wording l:] Commercial OR \ Typeface Land Use/Cover \ Neither; they are e . Legend Land Use/Cover E Urban Land Use D Forest Land Cover [3 Commercral Land Use C] Rural Land Use E Dunes Land Cover Lend Use/Cover 1:] urb-n [3 Forest [3 Commercial \ Arrangement of Legend Features OR \ Legend Location \ Neither; they are equally important 1-. \ inclusion Of a Legend Neatline OR A Sequencing of Legend Features COLLEGE FOOTBALL ATTENDENCE (Odebcr 1992) 9.9.QQQ. - mu um um CmLEGE FOOTBALL ATTENDENCE COLLEGE FOOTBALL ATTENDENCE (October 1992) 1 (Odobfl m) 9999C...) in. new mum mm \ Neither; they are equally important Which is more important? 4* Arrangement of Legend Features _ . LandUstoverru Urban Forest Commercial Rural Dunes OR \ Inclusion of a Legend Neatline Land Use/Cover Urban Foreet Commercial Rural Dunn \ Neither; they are equally important _ .‘v> . Land Use/Cover [:1 Urban :1 Forest [:1 Commercial I: Dunes D Rural 4 Land Use/Cover E 1:] :1 1:1 1: Urban Forest Commercial Rural Dunn Which is more important? \ Inclusion of a Legend Neatline OR \' Legend Size \ Neither; they are equally important Which is more important? ; Clarity of (all) Legend Wording OR \ Text Spacing e la to or a 15 I00 no more \ Neither; they are e . uaIl im oortant Which is more important? é Visual Proximity of Legend Features (how close the symbols, M descriptions, and title g are to each other) . OR \ Legend Size ‘0 El: [:1 121: Urban Forest Commercial Rural Dunea \ Neither; they are equally important Which is more important? \ Sequencing of Legend Features OR .\ Visual Proximity of Legend Features (how close the symbols, descriptions, and title are to each other) 14 Neither; they are e . People per square mile DIE] -- so 51 76 101 201 or lo 10 to r less 75 100 200 more People per square mile to 10 100 200 more People per square mile I231: 201 101 76 51 50 or In to to or more 200 IOO 75 less Pimple Def sauere mile % . D “an - - - . 201 101 76 51 o to to to more 200 100 75 Which is more important? \ Clarity of (all) Legend Wording OR \ Conciseness of (all) Legend Wording Lewd . mle 3195.23) 1:] 201wmep°°plewum~ C1 iormzoopauploP-rnm- D rewroopooelewum. I D 51w75peqaleweqrfl. \“ D maimpeoplepereqnl. L Neither; they are equally important e” \ tie} in is r Which is more important? Land Use/Cover [:I Wetlands Land Use/Cover ‘ :I Wetlands Lakes [:1 Wetlands I: / , :I Lakes [:l Meadow Lakes [2 \ TeXt Locat'on Meadow Brush Meadow I :1 Brush 1:] Grassland Brush I: :I Grassland Land Use/Cover Grassland]: DESI: SEE: W L M o o 45 A G $ $/ 8 o o 9a \ Text Angle Wetlands Lakes Meadow Brush \ Neither; they are equally important ich is \Tyl \ Type Style OR 9L Clarity of (all) Legend Wording \ Neither; the are eq . 9 "II is l \Tel \M We Which is more important? \_ Text Spacing \ Type Style Wetlands Lakes Meadow Brush Grassmnd Wetlands Lakes Meadow Brush Grasshnd Wetlands Lakes Meadow Brush Grassbnd VVeflands Lakes Meadow Brush GrassMnd \ Neither; they are equally important wetmnds Lakes hfleadow Brush G re 5 sl 3 n d Wetlands Lakes Meadow Brush Grassland W @ Lilla W the Lakes lwead@w Brush @mmshnd Incl Leg \Le 'lism Q Which is more important? ELEVATION (Ieel above sea level) \ Inclusion of a Legend Neatline OR ELEVATION (feet above sea level) / \ Legend Location ELEVATION ' \ (leel above aea level) :7 \ Neither; they are equally important \Leg \Lt 'ih is n \ Legend Size OR A Legend Location W \ Neither; they are equally important \Se .\ L£ People per SQLBI'B mile i/t _ rj’ _ ‘ -201ormore Lg Sequencmg of -10110200 Legend Features ‘ 1 OR \ Legend Shape \ Neither; they are equally im hhsn .Tex '\ Cla Lu \Ne G <4 69 E E 00,) To, /> 86‘ $/ \ Text Angle UEOUzemg 13:: Wetlands Lakes Meadow Brush Land Use/Cover E Wetlands :Wetlands \ Clarity of (all) C) Lakes E Lakes Legend Wording E Meadow E Meadow :1 Brush 1:] Brush I: Grassland E Grassland 4 Neither; they are equally important ‘\ Vl Le .Lel ‘th Is I OR R Visual Proximity of Legend Features (how close the symbols, descriptions, and title are to each other) ‘th is 1 the} LT; Which is more important? Urban Forest fig Text Spacing Com mercial Rural Dunes Land Use/Cover [: Urban E Forest Urban Forest . Commemwl Rural Dunes Urban Forest Cornntercial R ural D un es Land Use/Cover Land Use/Cover Jam! UH/C’m, :l Urban E::]Forest :l Urban :1 Url“ E Forest :l jar-n! \ T f ype ace I: Commercial :l Commercial I: Commercial :l Comm-rm, E Rural E Dunes E Rural E Dunes \ Neither; they are equally important E: Rural l:l Pm, E Dunes E Earle! J: \T Which is more important? Land Use/Cover E Wetlands Land Use/Cover E Wetlands Lakes [: Wetlands E \ Text Location I: Lakes E Meadow Lakes E [: Meadow Brush l:l Meadow E :] Brush E Grassland Brush 1:] E Grassland Land Use/Cover Grassland: l: l:l E E Wetlands Lakes Meadow Brush E E E l: Wetlands Lakes Meadow Brush SEE: Wetlands Lakes Meadow Brush \ Type Size L Neither; they are equally important “this: \Ty Which is more important? Crop Suitability Crop Sui tabl iCY Crop Suitability 6,0,, Sui/glib, E Very High I: Very High :] Very High [:1 V", J!.',/: \ Typeface l: High C] High I l: JUI'M E Medium :1 Medium ' [:1 777.1.” :Low [:1Low [:ILow :10.” l: Not Suitable E Not Suitable :] Not Suitable [:1 MolSuilal/n EEC]: EEC]: M < H L L ’9. Q / \ Text Angle lG \C/)V 0'13, 06 H ((96 47 80,. 01,, ‘42; :1 :l l:l [:1 Very High High Medium Low L. Neither; they are equally important OR J; Arrangement of Legend Features ‘lch is \ Visual Proximity of, Legend Features (how close the symbols descriptions, and title are to each other) OR \ Arrangement of Legend Features L Neither; they are equally im -ortant Land Use/Cover Commercial Rural Dunes Urban E: Land Use/Cover \ Inclusion of a Legend Neatline \ Neither; they are equally important Which is more important? Land Use/Cover E Urban \ Conciseness of (all) E Legend Wording 1:, 22::ercia. :1 Rural E Dunes OR \ Text Angle QC,— Neither; they are equally important Legend Land Use/Cover E Urban land use E Forest land cover I: Commercial land use E Rural land use E Dunes land cover 3 o f6$ E E o '9 008 o, '9/ ’) \S‘ Dunes Rural Forest Which is more important? ONE DOT EQUALS 10 PEOPLE O \ Legend Location ONE DOT EOUALS1O PEOPLE O \ Sequencing of Legend Features OR \ Arrangement of Legend Features XL Neither; Peoplepersquarem‘le - 201 ormore - 10110200 - 201ormore Peoplepersqa'errile they are equally important \L Which is more important? \ Legend Location People per square mile l: E 50 51 101 201 or to c or less 75 100 200 more OR A Sequencing of Legend Features . I People per square mile or lo to to or less 75 100 200 more \ Neither; they are equally important People per square mile :l---- 50 51 75 101 201 Of lo [0 0 of less 75 100 200 more People per square mile 201 101 or to to to more 200 100 75 ---l: 51 so or less Which is more important? l: E E :l UrbanDunesRuralForest ESE: #TYPeSV-e Urban Dunes Rural Forest ESE: Urban Dunes Rural Forest Land Use/Cover Land Use/Cover Land Use/Cover Jam! Use/Cauor E Urban :1 Urban l:l Urban 2 Hr!“ \ Typeface E Fore“ E Forest E Forest :1 5“”; E Commercial E Commercial E Commercial I: Commercial :l Rural [2 Rural l:l Rural l:l Pam! I: Dunes E Dunes E Dunes l: 25“.: \ Neither; they are equally important | I. \ . I I l . \ b . ' Which is more important? Crop Suitablity I: Very High E Very High High :l LT . EHigh EMedium ext Location S Medium Low [:1 2 Low E Not Suitable E NOL Suitable Crop Suitability Very High Very High Very High High High High \Type Style Medium Medium Medium Low Low Low Not Suitable Not Suitable Not Suitable \ Neither; they are equally important Crop Suitability Very High E High :1 Medium [: Low Not Suitable E Wary High ifii iy )9 all] (a) did Mr [L a W M 62: ii S trim/hie 4 C '2: Which is more important? Very High Very High Very High \ Type Style ngh' High High Medium Medium Medium Low Low Low NotSuitable NotSu/tab/e Not Suitable E High E High :1 Medium :1 Medium :1 Low :3 Low «4 Type Size \ Neither; they are equally important Very High [it ad] , f; (2 w M rm 2? 52?; J] it‘s-J if.) he Very High Q Very High 2 very High E ‘ich is \Cl \T \/ Clarity of (all) egend Wording OR \ Type Size \ Neither; they are equally important Which is more important? Wetlands Wet/ands Lakes Lakes \Type Style Meadow Meadow Brush Brush Grassland Grass/and OR Wetlands Lakes Meadow Brush Grassland Land Use/Cover Land Use/Cover Land Use/Cover JanJZIAI/C’m, :l W etlands E Wet lands l: Wetlands E MAI/“J, T eface I I Lakes I lLakes VP [:1 Meadow D Meadow C] Brush l: Brush E Lakes E Ja‘u |:l Meadow E ”7me 2 Brush [2 grue‘ E Grassland l: Grassland E Grassland I: gnu/anal \ Neither; they are equally important \T Which is more important? 3 Verz High E Very High Q VeryHigh E \ Type Size S H lg [:1 High C High Q :Medlum :Medlum :Medium u E LOW :1 Low [:1 L... \ Text Angle Very High High Medium Low L Neither; they are equally important Which is more important? U ‘D U U 0D e9®D «fig 0) O I EC)"- 9? be ‘9. \ Text Angle E :1 E 2 Very High High Medium Low ICU—I -