Using Statistical Figures to Present Research Findings
Statistical figures transform results into visual patterns that readers can compare quickly. A good figure answers a defined question accurately, labels units and denominators, displays uncertainty where relevant and avoids visual distortion.
Learning objectives
- Select figures for categorical, continuous and time-based data.
- Construct clear tables, bar charts, histograms, line graphs, scatter plots and box plots.
- Present uncertainty and subgroup comparisons.
- Identify misleading graphical practices.
Choosing the display
| Purpose | Preferred display | Example |
|---|---|---|
| Compare categories | Bar chart | Vaccination coverage by district |
| Show continuous distribution | Histogram or box plot | Patient waiting time |
| Show change over time | Line graph | Monthly malaria cases |
| Examine two numerical variables | Scatter plot | Age and systolic pressure |
| Show effect estimates | Forest plot | Risk ratios with 95% CIs |
| Detailed exact values | Table | Baseline characteristics |
Figure construction checklist
- Self-explanatory title stating what, who, where and when.
- Clearly labelled axes, units, categories and legend.
- Appropriate scale and equal intervals.
- Sample size or denominator where interpretation requires it.
- Confidence intervals or error bars labelled correctly.
- Source note for adapted or secondary data.
- Readable colours, patterns and font, including grayscale and colour-blind accessibility.
Worked scenarios
Monthly outbreaks
Use a line graph because order and time trends matter. Mark interventions without implying they caused changes unless design supports it.
Length of stay
Use a histogram for shape or box plots for group comparison. A bar chart of means alone can hide severe skew and outliers.
Tables
Number tables in order, give each a concise title, place variables in rows and comparison groups in columns, report units and define abbreviations in notes. Avoid repeating every table value in the text; describe the important pattern.
Misleading presentation
- Truncated axes that exaggerate differences.
- Three-dimensional effects that distort area or height.
- Unequal time intervals displayed as equal.
- Pie charts with many categories or similar slices.
- Changing denominators across bars without explanation.
- Error bars without stating whether they show SD, SE or CI.
- Selective omission of inconvenient groups or time points.
Interpretation example
A chart shows coverage rising from 82% to 86%. A zero-based axis reveals a modest four-percentage-point increase; a truncated 80–87% axis may make it appear dramatic. Report both the absolute change and the denominator, and assess uncertainty before claiming meaningful improvement.
Figure versus text
Use text for one or two central numbers, a table when exact values matter, and a figure when the pattern or comparison is the message. Do not present identical data in all three formats.
Review questions
- Choose displays for prevalence by region, age distribution and monthly cases.
- Explain why histograms and bar charts are different.
- Identify four misleading graph practices.
- Design a table of participant characteristics.
References
- WHO. Health Research Methodology.
- CDC. Principles of Epidemiology in Public Health Practice.
Educational content only; follow the reporting standard required for your study design.
