4.4.1 Designing Tables and Figures: The Visual Presentation of Information

Tables and figures may, of course, appear in other parts of a thesis (in a background or methodology chapter, for instance), but they commonly play an essential role in the chapter(s) dedicated to reporting the results of doctoral research, so I have chosen to discuss them here. As I mentioned in Section 1.3, tables and figures can be embedded in appropriate places within the text or tacked on to the end of the relevant chapters or to the end of the thesis as a whole. The size and structure of tables and figures, as well as committee preferences and university or department guidelines may influence placement decisions. Your university or department might even have specific guidelines regarding the preparation of tables and figures for doctoral theses, or there may be standard canonical forms for tables and figures within your discipline, so do discuss with your supervisor the inclusion and design of any tables and figures you are planning. It is also a good idea to give some thought to the software you are using to construct your tables and figures: if, for instance, you will be designing your tables and figures in the same programme that you use to write the text of your thesis, there should be no problems with embedding tables and figures in your text, but if you are using a different programme for creating tables and figures than you are for creating text (or even two additional programmes – one for tables and the other for figures), it is important to make sure that all the software you use is compatible and will therefore allow you to combine text, tables and figures as necessary. 

As a general rule, all tables and figures should be numbered, usually with Arabic numerals (Roman numerals or letters are much rarer), in the order in which they are first mentioned in the thesis, and each table or figure should be referred to in the thesis by its number (e.g., ‘Table 1’ and ‘Figure 2’) along with some explanation of what the reader should look for in the table or figure. All tables or all figures may be numbered in a single sequence throughout the thesis, or the numbering may begin again for each new chapter; if the latter is the case, the chapter numbers will also need to be provided before the table and figure numbers (e.g., ‘Table 3.1’ for the first table in Chapter 3 and ‘Figure 4.2’ for the second figure in Chapter 4). The tables and figures themselves should also appear in their order of mention, whether they are embedded in the text or placed at the end of the chapters or the thesis as a whole (if you will be including tables and figures in appendices at the end of your thesis, see Section 4.6.1 for advice on numbering and labelling such tables and figures appropriately). If tables and figures are embedded, they should be placed as close as possible to your discussion of or references to them in the main text. Neither tables nor figures should serve as substitutes for describing methodology, for presenting results or, more generally, for scholarly argumentation, but both should augment and clarify your methods, findings and argument. Each table or figure should ideally be designed so that it can stand alone, which means that all of the information in it should be explained or defined in the context of the table or figure itself so that the material contained within it is entirely accessible to readers without recourse to other parts of the thesis.

Tables. Although the line distinguishing tables from figures (and even lists, on which see Section 5.5.2) can be somewhat blurry at times, generally speaking, tables show numerical and/or textual information in neatly arranged rows and columns that render the material visually accessible and allow it to be more easily compared, calculated and understood by readers than it could be were the same data simply described in text. The overall layout of a table, all the elements contained within it and the heading and notes that appear above and below it must be clear, concise, informative and accurate. This means that a table must use fonts, characters and symbols in sizes and styles that are highly legible and readily comprehensible (for some excellent examples of table formats, see the Publication Manual of the APA, 2010, pp.129–149; Tufte, 2001 provides guidance on designing tables that may also prove helpful). You may want to take advantage of automatic table functions in the software you are using to construct your tables, or you may prefer a more manual approach using tabs and spaces; your department or supervisor may have advice on which method is preferable. Rules (or lines) can be helpful for separating and clarifying the information presented in tables, and most tables are demarcated by head and tail (top and bottom) rules (note that the table heading should appear above the head rule and the table footnotes beneath the tail rule), but vertical rules are often frowned upon and even horizontal rules should be kept to a minimum. The length and position of any rules used should be appropriate so that they effectively separate and clarify the material presented.

Each table included in a thesis should bear at its top a title or main heading that identifies the table by number and indicates exactly what the table shows; any repetition of the heading in the text of the thesis or in a list of tables (if included: see Section 1.1.9 and Section 4.6.2) should match that heading exactly. A table heading should be as short as possible, but detailed enough to explain briefly what the table contains and shows. ‘It should not furnish background information, repeat the column heads, or describe the results illustrated by the table. . . . Let the table give the facts; commentary can be offered in the text’ (Chicago Manual of Style, 2003, p.499). A full stop usually appears between the number and the heading – ‘Table 1. Demographic characteristics of study participants’ – but occasionally a colon is used instead, and no punctuation at all is quite common, especially if the table bears a double number (Table 2.1 Demographic characteristics of study participants), in which case an extra space (or two) is sometimes added after the last numeral. Only the initial letter of the first word of the heading is usually capitalised (as in the preceding sentence), along with any proper nouns, but alternatively all main words can bear capitals: ‘Table 1. Demographic Characteristics of Study Participants.’ Abbreviations should be avoided in table headings (as in all headings: see Section 6.1, and on abbreviations, Section 6.3) if at all possible, but if any nonstandard abbreviations also used in the table must appear in its heading, they can be defined parenthetically: ‘Table 1. Demographic characteristics of SPs (study participants)’ would be effective, for instance, though most nonstandard abbreviations in a table should be defined in the table footnotes (see my discussion below regarding abbreviations and footnotes in tables). Units that are used throughout a table can also be defined or explained in its heading – ‘Reaction time of study participants (in milliseconds)’ – and the logic governing the order in which material is presented in a table can also be briefly explained in the heading: ‘Spring wild flowers of the Fraser Valley (arranged according to blooming times).’ No full stop is required at the end of short table headings, but if a table heading runs to more than one sentence, a closing stop is appropriate, and if this is the case for one or more table headings, a full stop can be used to close all table headings so that consistency is maintained across all tables in the thesis.

Each column and row of a table should bear a heading that indicates the kind of data contained in the column or row. Column and row (or stub) headings should be clear, precise and legible, with all of them ideally using similar phrasing or syntax (when relevant) and appearing in the same size font (often a font slightly smaller than that used in the thesis). Since space is usually limited in a table, these headings should also be as brief as possible, should feature initial capitals on their first word and proper nouns only and should use no final punctuation. Units and percentages used in the column or row must be identified in the relevant heading (unless they are common throughout the table and mentioned in the main table heading), which is often done in parentheses: ‘Reaction time (ms).’ Parentheses can also be used to present more than one kind of data: for example, a column with the heading ‘No. of men (%)’ might contain entries such as ‘14(28),’ ‘24(48)’ and ‘32(64),’ with the first numeral representing the number of men, the second numeral (in parentheses) providing the percentage and both doing so clearly and concisely without the need to repeat the units. Units should not be repeated in every cell of a column or row unless the units differ in each case and providing a single unit for the whole column or row would be inaccurate. As the above examples show, abbreviations and symbols can be used in column and row headings (see my discussion of abbreviations in tables below), but column and row headings are not normally numbered unless numbers are used in the text of the thesis in relation to the material presented in the columns and rows, in which case the numbers can be helpful for readers using the table.

Numbers in tables should in almost all cases appear as numerals, not words, which allows for easy calculation and comparison as well as saving space. Any mathematical operators (‘+,’ ‘-,’ ‘>’ etc.) should be placed up against the numerals they apply to, and special care should be taken to ensure that minus signs are distinguished from hyphens and dashes. Spans of numbers that appear in column or row headings should be accurate with neither gaps nor overlaps: ‘1–10,’ ‘11–20,’ ‘21–30’ and so on rather than ‘1–9,’ ‘11–19,’ ‘21–29’ or ‘1–10,’ ‘10–20,’ ‘20–30.’ The numbers in each column of a table should also be aligned vertically in terms of decimal points, commas and spaces if the rows contain similar units and especially if the column has a total. Consistency in the format of numbers is essential within a table and best across all tables in a thesis, and accuracy in reporting numerical data is essential, so make sure that all numbers in each table you include are correct and match exactly the same numerical information elsewhere in the thesis (for more information on using and formatting numbers, see Section 6.4). If tables are to be compared, the same units should be used in all of them, and wherever the same units are used, the format of those units (terms, abbreviations etc.) should be identical.

Abbreviations and symbols save space, so they are used extensively in tables, but they are only effective if readers understand what is meant by them, so common abbreviations and symbols for units of measurement (e.g., ‘°C,’ ‘Hz,’ ‘g,’ ‘kg,’ ‘m’ and ‘km’) and standard statistical abbreviations and symbols (‘ANOVA,’ ‘ES,’ ‘N,’ ‘p,’ ‘RMSEA,’ ‘SEM’ etc.) should be used whenever possible. These abbreviations usually do not require any definition, though if there is any doubt about your readers’ familiarity with the statistical abbreviations in particular, definitions should be provided either in the main table heading (see above) or, more likely, in the table footnotes (see below); abbreviations are usually not defined in column or row headings. When any nonstandard, unusual or specialist abbreviations are used, such as those for highly technical or discipline-specific terms or those for group names and other aspects of a particular study, they must be defined in the table even if they have already been defined in the thesis (see Section 6.3 for further information). If such an abbreviation appears in both the table heading and the table itself, it can be defined in the heading, but otherwise, abbreviations of this kind are usually defined in the table footnotes (see my discussion below). Whatever kinds of abbreviations are used, they should be used in the same forms in all relevant tables as well as in other parts of the thesis, so that consistency and clarity are maintained.

Generally speaking, there are four different kinds of footnotes that can appear, as needed, at the bottom of a table: general notes, source notes, notes on specific parts of the table and probability notes. Table footnotes are often set at table width in a slightly smaller font size than that used in the table itself, which, when a table is embedded, helps to distinguish them from the running text of the thesis beneath the table, but a line or two of space should nonetheless appear beneath the table footnotes for a clear layout. Strictly speaking, each note should begin on a new line and end with a full point, but this is often impractical for reasons of space and layout, so shorter notes of a single type can run on separated by semicolons – ‘ANOVA = analysis of variance; CI = confidence interval; ES = effect size’ – and longer notes that differ more greatly from each other but are still in the same category sometimes run on separated by full stops: ‘Data were collected from 1 May 2003 to 30 April 2004. ANOVA = analysis of variance; CI = confidence interval’ and so on. There is no need to use any type of note unless that type of explanation or documentation is required, but the types of notes should always be arranged within a single thesis in the same order under each table that uses more than one type. There is, however, no uniform agreement on what that order should be or exactly what should be contained in each type of note (though probability notes are a little more straightforward than the others), so do check department guidelines for relevant advice or discuss the matter with your supervisor, and if you find you need further help designing your table notes, more details can be found in Butcher et al. (2006, p.226), the Chicago Manual of Style (2003, pp.511–513), the Publication Manual of the APA (2010, Section 5.16) and Ritter (2005, Section 15.2.5). Whatever order and content you decide on for the four categories, consistency should be observed across all tables in a thesis, and any information in the footnotes repeated elsewhere in the thesis should, of course, correspond with precision.

I outline below the most common and accepted scholarly practices for constructing the four kinds of table footnotes:

• General notes, as their name suggests, apply to the table generally or as a whole and usually appear first of the four types, though sometimes source notes precede them or are included in the general notes. General notes usually begin with the word Note or Notes (often in italic font as here, although bold font or block capitals are also used), which is followed by either a full stop or a colon; font and punctuation should be consistent throughout the tables in a thesis. No indicator linking the general notes category or the individual notes within it to any part of the table is required. Definitions of abbreviations are often included in the general footnotes, so ‘Data were collected from 1 May 2003 to 30 April 2004. ANOVA = analysis of variance; CI = confidence interval; ES = effect size’ would work as a general note. When abbreviations are included in a general note in this way, they are usually recorded in alphabetical order just as they are in a list of abbreviations (see Section 1.1.7 and Section 4.6.2).
• Source notes often appear as a separate category rather than being included in the general notes, which they can follow or precede. They usually open with the word Source or Sources (normally set in italic font, although here, too, bold font or block capitals are sometimes used), which is followed by either a full stop or a colon; the format should match that of the Note or Notes preceding the general notes if there are any and remain consistent across all tables. Source notes provide references for any material (the entire table, its format, the data within it etc.) that has been borrowed from another work. Sources can be provided via whatever referencing system is used in the thesis as long as each source is included in the list of references or bibliography; otherwise, a full bibliographical reference should be provided, with this format being particularly appropriate for manuscripts, visual arts and other sources for which recording the library or museum and its location alongside the material is standard practice (e.g., ‘Based on Washington, DC, Folger Shakespeare Library MS V.a.354, fol.121v’).
• Notes on specific parts of tables usually follow general and source notes and are like true footnotes in that they are keyed to specific words, symbols and numbers within the table. To avoid confusion, the system of indicators used in a table must be separate and different from that used for any footnotes or endnotes used in the main body of the thesis, but the same system should be used (wholly or partially as relevant) in all tables requiring notes on specific parts. Arabic numerals can therefore be used if numbered notes are not used elsewhere in the thesis, but if there are already numbered notes in the main text, it is best if the table notes take another form, such as superscript lowercase letters or a collection of specific characters; it is essential that these cannot be confused with other characters and symbols in the table(s), however, and the asterisk (*) should definitely not be used if probability notes are also required. A table should be read across the rows from top left to bottom right when providing indicators within the table and arranging the specific notes at the bottom. Such notes can be attached to any element of a table requiring explanation except the main table heading (footnotes are usually discouraged on all main headings) and probability and significance levels.
• Probability notes follow all other categories. Like notes on specific parts of a table, they use a system of indicators, but one that specifies p values via the number of asterisks attached to a numeral and is generally explained with a note of this kind: ‘*p < .05; **p < .01; ***p < .001’ (the Publication Manual of the APA recommends that any value smaller than ***p < .001 not be used: 2010, p.139). Note that the ‘p’ is usually lowercase and set in italic font, zeroes are normally omitted before the decimal point and spaces generally appear around the ‘<’ symbol, though all three of these details can vary as long as consistency is maintained for each element throughout a thesis. As with the notes on specific parts of a table, probability notes should be added by reading along the rows of a table from left to right and top to bottom, and whatever system and precise formats are applied in one table should also be used in all other relevant tables included in the thesis.

Figures. Many different illustrative elements can be included as figures in a thesis, including charts, graphs, plots, boxes, photographs, paintings, drawings and maps. Figures should not be included for aesthetic purposes only, however; they should enhance, clarify and/or facilitate the presentation of information, adding substantively to the argument of the thesis in a way text alone could not, just as tables should. Much of what I say about tables above applies also to figures, but for more detailed treatments of illustrations and figures, see Butcher et al. (2006, Chapter 4), the Chicago Manual of Style (2003, Chapter 12), the Publication Manual of the APA (2010, pp.150–166) and Ritter (2005, Chapter 16). Tufte’s classic guide (2001) to designing statistical graphs and charts may also prove helpful, and Monmonier (1993) provides specific advice on maps. Generally speaking, when considering and/or designing figures for inclusion in your thesis you should ensure that each figure will actually achieve for your reader what you would have it achieve, and several aspects of figures must be carefully considered to guarantee that this will be the case.

Photographs, for instance, must be clear and represent accurately the reality they are intended to depict. Any changes you need to make to photographs, especially those purchased from libraries, museums and the like, should be acknowledged, and the identities of any subjects that appear within them may need to be masked. Charts and graphs should be complete, with lines, bars and shading effectively showing what they should, all scales and units defined and clearly indicated and the necessary axes marked. Maps should show the areas and locations discussed in the thesis and all place names should match the names for the same places used in the main text. Indeed, all information contained in figures of any kind should be accurate in relation to the same information presented elsewhere in the thesis, and figures should not contain extraneous material that might distract the reader from the primary purposes of the figures. Similar material presented in more than one chart, graph, map or any other kind of figure should, as much as possible, use consistent design elements, graphic styles, typography and terminology.

Unlike tables, figures do not have main headings; instead, captions and/or legends are used. The terms ‘caption’ and ‘legend’ can be confusing, however. ‘Caption’ usually refers to what is basically the equivalent of a table’s main heading – a brief and often explanatory description of what the figure shows that also includes the word ‘Figure’ (or the abbreviation ‘Fig.’) and the figure number. The caption can help guide the reader through understanding different parts or elements of the figure via instructions such as ‘to the left,’ ‘at top right,’ ‘in the bottom image’ and so on, or via letters (‘a,’ ‘b,’ ‘c’ etc. or ‘A,’ ‘B,’ ‘C’ etc.) that label the different parts and are explained in the caption. A figure caption usually appears below or sometimes beside the figure it describes, very rarely above it, but figure captions can and usually should use the same style and format (wording, font, capitalisation, punctuation etc.) as the table headings in a thesis do (e.g., ‘Figure 1. Letter forms in the three British Library manuscripts,’ which matches the format of ‘Table 1. Demographic characteristics of study participants’ that I used as an example above). ‘Legend’ is sometimes used interchangeably with ‘caption,’ but a legend tends to provide more detailed information or a more extensive explanation than a caption does. A legend can be an integral part of a figure (see the Publication Manual of the APA, 2010, p.159), usually appearing in the bottom part of the figure (but not beneath it as the caption does) and using the size and style of font used within the figure. With this meaning, a legend overlaps with a key, which also appears within the figure, defines the symbols and abbreviations used in the figure and explains any other aspects of the figure that might be unclear without description. More than one of these (caption, legend and key, that is) can be used while constructing figures, symbols and abbreviations can be defined in any one of them, and sources, credits and acknowledgements are usually provided at the end of the caption or legend. Whatever the exact formats and strategies you use for figures may be, it is essential that consistency in their use be maintained as much as possible throughout all figures in your thesis.

The labels in a figure are equivalent to the column and row headings in a table, and should be equally clear as well as both concise and descriptive. All important elements of a figure should be accurately and effectively labelled, with each label appearing as close as possible to the part of the figure it identifies, using a leader line to connect the label to what is labelled if necessary and bearing an initial capital only on the first word and any proper nouns (‘Top right’ and ‘Bottom left,’ but ‘British Columbia’). Although the same size and style of font can and often should be used for all labelling within a figure, font sizes and styles can be altered to distinguish different categories. In a map, for example, countries could be labelled with block capitals (CANADA) while the labels of other elements use initial capitalisation only, with regions or provinces marked in roman font (Alberta), cities in bold font (Victoria) and rivers in italic font (Thompson River). If font sizes vary, the variation should not be extreme (generally speaking, there should not be more than a 20% difference in font size within a single figure), and all fonts and individual labels must remain legible. Levels and axes should be labelled, and scales and units of measure provided. Abbreviations and symbols can be used, but as with those found in tables, anything nonstandard or potentially unfamiliar to the intended audience should be defined (see my next paragraph for further details). Excessive labelling can be avoided by using a long caption and/or a detailed key, which are often more legible and thus more successful tools for explaining a complex figure. As with the headings within tables, the style, format and content of labelling should, as much as possible, maintain consistency within each figure and across all relevant figures in a thesis, and they should also be consistent with any information repeated in other parts of the thesis.

Footnotes are not used in figures, so definitions of abbreviations and symbols, information on probability values, explanations of specific aspects of a figure and documentation of sources along with any necessary credits and acknowledgements are provided in the caption, legend or key, with some elements in one and some in another if more than one of these is used. Definitions of symbols or colours, for instance, might be more appropriate in a short key, while definitions of abbreviations would be better suited to a legend or caption. Guidelines differ on precisely how the different elements should be arranged, so do check with your department or supervisor, but general explanations tend to come first, followed by definitions of abbreviations and probability notes (using the same asterisk system as that used in tables) and almost always finishing with sources and acknowledgements (some good examples formatted as legends within figures can be found in the Publication Manual of the APA, 2010, pp.152–166). References can be provided via the system used elsewhere in the thesis as long as the sources are included in the bibliography or reference list, but full bibliographical information can instead be recorded in the caption or legend and this approach may be more appropriate for some material (see this point in the discussion of tables above). Longer notes and explanations are usually structured as full sentences, but shorter ones, such as definitions of abbreviations or asterisks for p values, can be presented as run-on lists with the items separated by semicolons (on lists, see Section 5.5.2 below). Abbreviations, symbols, definitions, explanations and their styles should remain consistent across all figures in a thesis and in all places where the same material is repeated in other parts of the thesis.

PRS Tip: The findings of your research constitute an extremely important part of your thesis, so it is essential to submit your results chapter(s) and your tables and figures (or even the initial sections of the results chapter and the first few tables and figures) to your supervisor before you proceed further with writing your thesis. The commentary of your supervisor and any other committee members you may wish to involve at this point can save you a great deal of rewriting in later parts of the thesis by bringing to light any problems with the ways in which you have chosen to report your findings, and thereby allowing you to resolve them at once by taking new directions if necessary. It is also a very good idea to have an objective and critical reader look over this material before you submit it to your supervisor and other committee members. Fellow postgraduate students make a good audience, but a professional proofreader has a lot to offer as well. The educated proofreaders at PRS not only specialise in academic and scientific writing, but some of them work particularly on theses and dissertations, and most of these proofreaders have successfully completed their own theses and/or dissertations. They can check and correct your spelling, punctuation and grammar as well as ensuring that your written English is communicating clearly with your readers in an acceptable scholarly fashion. They can examine tables and figures (and the references to them in your text) to be sure they are clear, consistent and effective for presenting the information they contain. They can check abbreviations, numbers, quotations, headings, fonts, references and a host of other complicated details for accuracy and consistency. They can, in short, produce results with which you will be delighted, so do send a section or chapter of your thesis to PRS today and discover the difference a professional proofreader can make.