Communicating results with scientific graphs

Graphs are great visual communication tools that, when used correctly, can consolidate large amounts of data to help identify patterns and relationships for an audience. Whether they are included as part of a scientific article, a presentation, or a poster, scientific graphs should help you to communicate the key messages or findings of your investigation.

Before you create a graph you should consider three things:

1. Do you need a graph?
Sometimes results can be easily summarised in a sentence or two, or by using a simple table. If you have a large number of categories with a variety of measurements, a table may be more appropriate to neatly display results.

2. What types of variables do you have?
Knowing the types of variables in your data and the statistical analysis you have performed will guide you when deciding what type of graph to use.

3. What is your message?
Graphs should clearly communicate a message to your audience. Therefore, you should only include data that will help communicate your message (while ensuring you’re not misleading your audience). You should keep this message in mind when creating and formatting your graph.

Graphs share common features that help your audience to understand your key message.

As a general rule, you should ensure that all of your figures for scientific articles or lab reports can be easily interpreted when printed in black and white. Colour can be used if your audience is likely to view the graph in colour (i.e. when viewing a poster or presentation) or if it essential to communicate your message.

How to display your data

Pie charts are rarely used in scientific articles, but they can be useful when communicating with the public. You should check the requirements of your assignment with your lecturer for guidance on how to display your data.

Types of variables

Continuous variable: Continuous variables are numeric measurements or observations that can include any number of values within a certain range. (E.g. temperature, time, weight, and concentration).

Discrete variable: Discrete variables are measured as whole units. (E.g. number of birds in a population, number of students in a class).

Categorical variable: Categorical variables describe a quality or a characteristic (E.g. Colour, species, sex, blood type).

Independent variable: The independent variable is the variable which you control or manipulate in your experiment, or the variable that you think will affect the dependent variable. Independent variables are placed on the x-axis of a graph.

Dependent or response variable: The dependent variable is the variable you think will be influenced by the independent variable. Changes in the dependent variable are observed or measured in relation to changes in the independent variable.

Example:

An experiment investigating the effect of light exposure on the rate of growth of a plant.

The independent variable is the amount of light exposure and the dependent variable is the rate of growth.

Should I use a table?

Sometimes a table will be more appropriate for displaying your data. Tables are great for displaying multiple variables, specific values, and comparing categories. A table will often require an audience to look up specific information to understand the data. Therefore, you should ensure your table is presented in a neat and logical manner.

It is important that you don’t just add all of your raw data to your tables. Similar to graphs, you need to consider the message in your data that you want to communicate to your audience. You may need to perform a statistical analysis on your data or summarise your results before adding the information to a table.

For large tables, you may need to shade alternate rows or highlight important details by using a bold font to allow your audience to read the table efficiently.

Figure & table legends

All of the tables and graphs that you create for scientific articles and lab reports will require a legend.

What is the point of a legend?

Identifies the graph or table (E.g. Figure 1. or Table 1.)
Informs your audience what the graph/table is showing
Provides any information that may be needed to interpret the graph/table
Some figure legends may need to include information on specific symbols or shading, the experimental methods, the error bars, or the sample size. For tables this information can be included in column/row headings or as footnotes.

The concise description in a table or figure legend should convey the key message of the table or graph to your audience without having to read the full article.

This module focuses on graphs and tables for use in scientific articles and lab reports. If you are designing a graph for a presentation or poster, you should refer to the relevant module for further design guidelines.

For detailed guidelines on creating figure legends, view the optional extras section below.

The essentials

Figure number (Figure 1 or Fig. 1)
You must refer to all figures in the main text of your report or essay. The figure number is used to allow your audience to find the figure you have referred to in your text.

Figure title
Figure titles can be descriptive or assertive.

A descriptive figure title briefly describes what the figure is displaying but lets the reader identify any trends or relationships, or is guided by the text you include in the results section. An assertive title can be used to identify a specific trend found in a graph or highlight the key message of a diagram.

Assertive titles can help your audience to quickly identify the key message contained within your figure but you should ensure your title does not mislead your audience or overstate your results.

Example 1.
Descriptive: Figure 1. Effects of dam construction on fish biodiversity.
Assertive: Figure 1. Dam construction results in loss of fish biodiversity.

Example 2.
Descriptive: Figure 2. Height distribution of two Eucalyptus grandis plantations in Queensland.
Assertive: Figure 2. Insect defoliation of Eucalyptus grandis reduces canopy height.

Figure legend examples

Figure legend – optional extras

Including any of the following optional extras will depend on what is displayed in the figure and what you feel your audience needs to understand the figure. The optional extras you include will also depend on what information you have included in your methods and how you refer to the figure in your results section.

View the examples above to see how the optional extras are used to describe a variety of figures.

You should check the requirements of your assignment or discipline for guidance on which optional extras to include in your legend.

1. Symbols, lines, colours & acronyms

If you have used symbols, lines, colours or acronyms in your figure that have not been identified on the actual figure, you need to ensure they are referred to in the figure legend. If you have used colour in your figure, make sure your audience will be able to view it in colour, otherwise the figure will be difficult to interpret.

Example
Oxygen consumption rate for Fish species 1 (filled circle) and Fish species 2 (hollow circle).

2. Means and error bars

If you are plotting mean values and including error bars, you need to state this in the figure legend.

Example
Mean trunk diameter (+\- SEM) of Eucalyptus grandis.

3. Statistical information

Some figure legends will mention the type of statistical test used, the sample size, p-values, or other statistical information. The inclusion of this type of information often depends on personal preference or editorial guidelines.

However, it can be useful to include this type of information in figure legends to help communicate the validity of your results to your audience.

Example
Statistical analyses were performed using ANOVA with a Tukey’s post-test (***, p<0.0001; **, p<0.001; *, p<0.05; ns, not significant).

The sample size is often included in a figure legend when comparing two or more groups with varying sample sizes.

Example
Height distribution of two Eucalyptus grandis plantations in Queensland. Western ridge N = 350, Eastern ridge N = 300.

Whether you decide to include this type of statistical information in your figure legend or not, you must ensure it is included the text of your methods and results section.

4. Experimental information

Including specific experimental information in your figure legend can help your audience to distinguish between groups included in your figure.

For example, if you have labelled 3 different treatment groups using abbreviations, you should include more information on how the treatment groups vary. Obviously this information will be available in your methods sections but it will help your audience to understand the figure.

Example
The control (pH 5.3) was normal city tap water. The pH 3.5 and 2.0 water was acidified with 2 M sulfuric / 1 M nitric acid solution.

5. Compound figures

Compound figures can be used to display multiple related graphs or diagrams. Compound figures are useful when comparing results that can’t be contained within a single graph or diagram.

Compound figure legends still have a single figure label and title but each individual figure should be labelled (A, B, C, etc.), with a brief description of each provided in the figure legend.

6. Referencing

If you have obtained a figure from another source (rather than creating your own) you must create your own figure legend and cite the author.

Example
Agricultural water use, by state 2013-14. Reproduced from Australian Bureau of Statistics (2015, p.18).

If you modify a figure, you should mention that it has been adapted from the source.

Example
Agricultural water use, by state 2013-14. Adapted from Australian Bureau of Statistics (2015, p.18).

Ensure that any citations in figure legends match the referencing style that you have used throughout your document. Any sources cited in figure legends must be included in your references section.

An Academic Explains

Useful links

Choosing the right graph
Extreme Presentation
Download PDF

Statistics for biology and agricultural science
by Ploughing Through Biometry
Visit website

Real chart rules to follow
by Flowing Data
Visit website