Dissertation Data Visualization: Charts, Graphs & Figures Guide

Why Your Results Chapter Needs Good Visuals

Data visualization in a dissertation is more than decoration — it’s how your readers understand what your analysis actually found. When you write 80,000 words of methodology and results, your reader’s brain is overwhelmed by text. A single well-designed chart can convey what half a paragraph cannot.

But here’s the problem: most students default to Microsoft Excel or SPSS default styling, which produces low-quality visuals. Pie charts dominate results chapters they shouldn’t. Truncated axes mislead reviewers. And most students have never heard of Edward Tufte’s data-ink ratio principle, which should be the bedrock of every figure they create.

This guide covers everything you need to select the right chart, format it according to APA 7th edition, avoid the most common mistakes students make, and use free tools to create publication-quality figures. You’ll walk away with a chart selection matrix you can actually use.

• Use position-based charts over angle-based ones. Cleveland & McGill (1985) showed that position along a common scale is far easier for the human eye to perceive than angle, area, or color — which is why bar charts consistently beat pie charts and bubble charts.
• One non-textual element per 1,000 words. Divecha et al. (2023) recommend this as a benchmark for manuscript balance. For a PhD dissertation at 80,000–120,000 words, plan for roughly 80–120 figures and tables in total.
• Limit pie charts to five slices or fewer. Nguyen et al. (2021) found the pie chart to be the single most misused chart type in scientific publications, with size violations accounting for 69.4% of all visualization pitfalls.
• Jamovi and JASP have replaced SPSS at many universities. Both are free, open-source, and produce APA-style output natively — jamovi reports over 300 university adoptions and 5 million downloads.
• One-third of visualization errors involve color. 39.7% of visualization mistakes stem from color violations. Use colorblind-safe palettes like viridis or ColorBrewed and verify contrast before submission.

Chart Selection Framework: The Cleveland & McGill Perceptual Hierarchy

The most important thing you need to know about choosing a chart type is that your human visual system can perceive certain visual channels far more accurately than others. Cleveland and McGill (1985) studied the visual characteristics of data visualization and ranked them from easiest to hardest for the human eye to perceive:

1. Position along a common scale (easiest)
2. Position along a non-aligned scale
3. Length
4. Angle and slope
5. Area
6. Volume, density, and color saturation
7. Color hue (hardest)

This ranking explains everything. Bar charts beat pie charts because bars use position and length — the two easiest visual channels to perceive. Pie charts use angle and area — channels the human eye struggles to compare accurately. Bubble charts add volume, which our brains interpret poorly. This is not a subjective preference — it’s cognitive psychology.

That’s why I recommend bar charts and box plots for most dissertation results chapters. They’re clearer, more defensible, and aligned with perceptual reality. If you’re using pie charts, you need a strong justification — and even then, keep them to five slices or fewer.

Here’s why the chart selection matters.

The Chart Selection Matrix

The table below maps data types to the most appropriate chart types, rated by Cleveland & McGill perceptual accuracy. Use this matrix as your starting point when designing figures for your results chapter.

Chart Type	Best For	Cleveland & McGill Rating	When to Avoid	Alternative
Bar Chart	Comparing values across discrete categories	Rank 1–2 (Position, Length)	Too many categories (>15)	Dot plot (cleaner for many categories)
Line Graph	Showing trends over time; continuous variables	Rank 1–3 (Position, Slope)	Categorical data with no temporal sequence	Stacked area chart for cumulative trends
Scatter Plot	Relationship between two continuous numeric variables	Rank 1 (Position)	Overlapping points with large datasets (N > 500)	Hexagonal binning or 2D density plot
Box Plot	Comparing distributions, medians, and outliers across groups	Rank 1–2 (Position, Length)	When readers are unfamiliar with box plot conventions	Violin chart (shows full distribution shape)
Histogram	Frequency distribution of continuous data	Rank 1–2 (Position, Length)	Categorical data (use bar chart instead)	Frequency polygon (connects midpoints)
Pie Chart	Part-to-whole relationship (≤5 categories only)	Rank 5–6 (Angle, Area) — avoid if possible	More than 5 categories; when exact values matter	Bar chart or donut chart (cleaner than pie)
Heat Map	Complex relationships across many dimensions	Rank 6–7 (Color hue, saturation)	Colorblind audiences; small print	Diverging palette with labeled cells
Dot Plot	Ranking or comparing many items	Rank 1–2 (Position, Length)	When bars convey stronger categorical grouping	Horizontal bar chart (easier to scan)

This matrix is the most practical takeaway from the perceptual hierarchy. If your research requires comparing categories, a bar chart or dot plot is almost always the right choice. If you’re examining relationships between variables, a scatter plot with regression line is the standard. If you’re showing distributions or group comparisons, a box plot is the gold standard.

A bar chart from UC Berkeley illustrates this principle well — position along a common scale makes the comparison almost effortless to read. The alternative, a pie chart with similar data, forces the reader to estimate angles and areas, which introduces perceptual error.

Quick Fact: The human eye can process five to seven well-chosen colors pre-attentively (Healey et al., 1995). That’s why qualitative color scales work best with three to five categories. Anything more requires the reader to constantly flip between legend and chart — cognitive load that slows comprehension and invites errors.

APA 7th Edition Formatting Rules

APA 7th edition is the dominant citation style in social sciences, education, and many graduate programs. Your results chapter figures must conform to its rules — and these rules are specific enough that most students get them wrong on first submission.

The Five Figure Components

According to APA 7th edition, every figure has five mandatory components:

1. Figure number — Bold, appears above the figure. Example: Figure 1 (no period after the number).
2. Title — Italic title case, one double-spaced line below the figure number. No period.
3. Image — The chart, graph, photograph, or illustration itself.
4. Legend — Positioned within the figure borders, explains any symbols used.
5. Note — Below the figure, flush left, non-italicized, beginning with Note. (italicized, period ending). Includes abbreviations, copyright attribution if adapted, and statistical details.

In your dissertation, refer to every figure by its number — no italics, capital “F”. Example: “As shown in Figure 1, the treatment group scored significantly higher.”

Practical Formatting Checklist

• Font size: 8–14 pt (depends on final figure size)
• Font: Consistent across all figures (Arial, Helvetica, or Times New Roman)
• Axis scales: Must start at zero (or explain if non-zero)
• No 3D effects: They add zero information and invite reviewer criticism
• Figure placement: Either all figures on separate pages after references, or embedded within text (check your program’s requirements)
• Labels: Every axis must have a clear label with units
• Color contrast: Avoid red/green combinations without checking colorblind accessibility

If you reproduce or adapt a figure from another source, you must include a copyright attribution in the figure note — plus a reference list entry. Permission statements are only required if you sought and obtained permission. For student coursework, attribution is still required but permission is not.

A practical tip: create your figures in the same tool you use for statistical analysis whenever possible. When jamovi or JASP generates a figure with APA-style output, you get the formatting right from the start — no manual reformatting required.

The 10 Most Common Visualization Mistakes Students Make

These mistakes are the single biggest reason students lose marks in their results chapter review. Each one is backed by peer-reviewed research on visualization pitfalls.

1. Using Pie Charts with More Than Five Slices

Nguyen et al. (2021) identified the pie chart as the single most misused chart type in scientific publications. Association rule mining showed that pie charts and size violations are the strongest predictors of misleading visuals (lift value 2.624).

The fix: If you have more than five categories, use a bar chart. If you need fewer than five, consider a donut chart (it reduces the same perceptual limitations while appearing more sophisticated).

2. Truncated Axes

Manipulating scales and axes is one of the most common mistakes in data visualization. Truncating the y-axis on a bar chart exaggerates differences that are actually minor. Dual-axis charts mislead when the scales aren’t comparable, implying correlations that don’t exist.

⚠️ Warning: Reviewers will penalize truncated axes in a dissertation. Always start your axes at zero unless you provide a clear explanation for why you can’t.

3. 3D Charts That Add Zero Information

3D charts are notorious for distorting perception. The depth dimension adds no new data — it merely makes reading the chart harder. Nguyen et al. (2021) found that high-dimensional representations used for low-dimensional data were among the most misleading visual formats students and researchers choose.

The fix: Remove all 3D effects. Use 2D bar charts, scatter plots, or line graphs.

4. Missing Error Bars

Error bars should be displayed on representative values in any chart comparing groups. Without them, reviewers cannot assess statistical significance visually. Nguyen et al. (2021) and Divecha et al. (2023) both emphasize that error bars (standard deviation, confidence intervals, or standard error) are essential for any graph showing group comparisons.

5. Cryptic Axis Labels

Axis labels should be short, clear, and indicate the measurement variable, units, and group sizes. Missing units or vague labels force readers to guess at meaning. This is one of the most common errors identified across all three major library guides (GW.edu, UMD, Berkeley).

6. Too Many Colors

39.7% of visualization errors involve color violations. Using too many colors or rainbow scales makes it difficult for readers — especially those with color vision deficiency — to distinguish categories. The human eye can only process five to seven well-chosen colors pre-attentively.

The fix: Use colorblind-safe palettes like viridis, colorblind-friendly from ColorBrewed, or tab10. Always verify contrast before submission.

7. Inconsistent Styling Across Figures

All graphs should be consistent in style and formatting. If Figure 1 uses a serif font and Figure 2 uses a sans-serif font, reviewers will notice — and it signals sloppy preparation.

8. Pie Charts Where Bar Charts Should Be Used

Shape violations — using the wrong chart type for the data — are the most common misleading visual format. Nguyen et al. (2021) found that bar charts used where pie charts should be used (and vice versa) were among the most frequent errors.

The fix: Use the chart selection matrix above. If you’re comparing categories, use bars. If you’re showing part-to-whole with ≤5 categories, consider a pie chart.

9. Overcrowded Scatter Plots

Scatter plots with large datasets (N > 500) suffer from overlapping dots. This is the single most common reason scatter plots fail to convey patterns.

The fix: Use hexagonal binning, 2D density plots, or reduce sample size. Many students don’t realize there are alternatives.

10. Missing Legends or Legends Outside Borders

APA 7th edition requires legends to be positioned within the borders of the figure. Placing legends outside the chart — while common in Excel — violates APA formatting rules and can lead to formatting rejection.

Pro Tip: Create figures in jamovi or JASP, not Excel. Both tools produce APA-compliant figures with legends inside borders, consistent fonts, and proper axis formatting by default.

Free Tools for Publication-Quality Figures

You don’t need SPSS or Tableau to create publication-quality dissertation figures. The following tools are free, open-source, and specifically recommended for student use.

jamovi (https://www.jamovi.org/)

jamovi is the single best tool for dissertation data visualization. It reports over 300 university adoptions and more than 5 million downloads. It has a point-and-click interface (no programming required), produces APA-style output natively, and generates figures with built-in error bars, proper legends, and consistent fonts.

Why I recommend it: If you’re doing frequentist statistics for your results chapter, jamovi covers t-tests, ANOVA, regression, and more — and exports every figure in APA format with one click.

JASP (https://jasp-stats.org/)

JASP is another excellent alternative to SPSS. It offers Bayesian statistics, frequentist methods, and machine learning — all with publication-ready figures. It’s particularly useful if your dissertation involves Bayesian analysis, which is increasingly common in social sciences.

RAWGraphs (https://app.rawgraphs.io/)

Need a Sankey diagram, sunburst chart, or beeswarm plot? RAWGraphs is a free web app that processes data locally in your browser. It’s perfect for complex, non-traditional visualizations that SPSS and jamovi don’t support natively.

BioIcons (https://bioicons.com/)

BioIcons provides free, customizable vector icons and diagrams for life sciences. If your dissertation involves biological data, this tool saves hours of figure design time.

draw.io (https://app.diagrams.net/)

For process diagrams, flowcharts, and methodology figures, draw.io is the go-to tool. It’s free, web-based, and integrates with Google Drive, GitHub, and WordPress.

Comparison note: If your dissertation already uses statistical software like SPSS or R, consider jamovi as a complementary tool for visualization. It doesn’t replace your analysis software — it produces better-looking figures with less effort.

Design Principles: Edward Tufte’s Data-Ink Ratio

Edward Tufte’s data-ink ratio principle is the single most important design concept you’ll never learn in a statistics course. From The Visual Display of Quantitative Information (1983, 2001), Tufte’s rule is simple:

Maximize meaningful information while reducing unnecessary complexity.

Every ink mark on your figure should convey data. Gridlines, background shading, decorative borders, and 3D effects — if they don’t add information, they add clutter.

Tufte coined the term “chartjunk” for visual elements that distract from the data. In a dissertation context, chartjunk is the single biggest reason figures get rejected by reviewers. Remove everything except the data, the axis labels, and the legend.

Practical Design Checklist

• Remove gridlines (or make them the palest possible, no lower than 15% tint)
• Eliminate background shading — use white backgrounds for highest contrast
• Remove default pre-styling formats from Excel or SPSS
• Minimize meaningless noise — maximize meaningful signals
• Keep font size legible at final print size (8–14 pt)
• Use a diverging palette for comparisons with positive/negative values
• Add error bars to any graph comparing groups
• Never use 3D effects — they serve no informational purpose

A bar chart from UC Berkeley demonstrates this principle: clean white background, simple black text, clear axis labels, no gridlines beyond the necessary ones, and a legend positioned inside the chart borders.

Figure Formatting Checklist

Copy this checklist and use it before submitting your results chapter. Every item corresponds to an APA 7th edition requirement or a common mistake students make.

Before Submission — Figure Review Checklist

• [ ] Figure number is bold and appears above the figure (no period)
• [ ] Title is italicized, title case, appears below figure number (no period)
• [ ] Legend is positioned within the figure borders (not outside)
• [ ] Note appears below the figure, flush left, italicized Note. (period ending)
• [ ] Axis scales start at zero (or explanation provided)
• [ ] Axis labels include units of measurement
• [ ] Error bars are included on group comparisons
• [ ] No 3D effects present in any figure
• [ ] Font size is 8–14 pt and legible at print size
• [ ] Color palette uses colorblind-safe options (viridis, ColorBrewed)
• [ ] Maximum of 7 colors per figure
• [ ] Styling is consistent across all figures (same font, same colors)
• [ ] No chartjunk — background is white or palest possible
• [ ] No pie charts with more than five slices
• [ ] No truncated axes without explanation
• [ ] If adapted from source: copyright attribution included in figure note

Summary: What to Do Next

Your results chapter is where readers spend the most time. Good visualization makes that time efficient. Bad visualization makes reviewers skeptical.

Here’s what I’d choose if I were designing your figures today:

1. Bar charts and box plots for category comparisons — they’re the clearest, most defensible charts available.
2. Scatter plots with regression lines for relationship analysis — the gold standard for continuous variable pairs.
3. Pie charts only when necessary — and only with five slices or fewer, preferably replaced by a bar chart.
4. jamovi or JASP for generating figures — both produce APA-compliant output with minimal effort.
5. Always apply Tufte’s data-ink ratio — remove everything that doesn’t convey data.

The charts, graphs, and figures in your dissertation are not decoration. They’re how your readers see what you found. Make them clear, make them APA-compliant, and make them error-free. If you do, your results chapter will stand out — not because of flashy visuals, but because your data will be readable.

If you’re struggling with your results chapter, remember: you don’t have to figure this out alone. Professional academic support exists — and the right help can turn weeks of frustrating figure design into a smooth, publication-ready process.

Related Guides

• Academic Writing Services — Dissertation writing support for all chapters
• Dissertation Guarantees — Quality standards and revision policies
• FAQ — Common questions about the dissertation process
• How It Works — Step-by-step support workflow

---

This guide synthesizes findings from peer-reviewed research on visualization pitfalls (Nguyen et al., 2021), APA 7th edition figure formatting rules (AUT, 2023), and library guides from UC Berkeley, George Washington University, and University of Maryland. All recommendations are evidence-based and designed for graduate-level dissertation work.