Figures and tables should not be used as a replacement for text, but to illustrate your findings or to help the reader understand complex data or display relationships between data. As such, they should only be included where necessary to inform the reader of information relevant to answering and enhancing the research question.
Use tables when exact numerical values are important. Use figures to show comparisons or trends. Decide on the most appropriate format between either table or figure and do not duplicate the same data in both formats.
Figures and tables should be self-explanatory and understandable through their title, body text and caption, without reference to the surrounding text.
Be aware of acceptable and unacceptable image editing. Journals are applying image manipulation detection tools could reject your manuscript if your images do not meet journal standards.
If you reproduce previously published material (such as a table, figure or photo), check its copyright status. If it was published in open access, check if the applied CC licence is more restrictive than the CC licence used by your target journal. If this is the case, the note about a more restrictive CC licence must be added to the caption/legend together with the source reference.
If it was not published open access, you will need to ask the copyright owner for permission. For many journals, this can be done via a copyright request button on the webpage of the paper, as in this example.
Use tables when it is important to show the absolute numerical values of your findings or when summarising large amounts of information (see example on next page).
Each column or row should have a heading, including units where applicable.
Choose the correct alignment for each column.
Omit data that are not informative (e.g. delete a column if it contains the same values in all rows – summarise in a footnote instead).
However, do not leave any cell in a table blank; use Not applicable, Not available, None detected, or similar to explain why the cell offers no data.
If the conducted research requires statistical analysis, be sure to present the results of statistical analysis in tables and indicate statistically significant differences.
Use abbreviations only if necessary for consistency or if there is not enough room for whole words. In captions or footnotes, define all abbreviations and symbols that are not obvious (e.g. error bars may denote standard deviation, standard error or confidence intervals).
Table example reproduced from Mihaylova et al 2023. Characterization of muffins reformulated with chia and lyophilized peach powder in terms of some technological and sensory aspects. Food Technol Biotechnol. 2023;61(3):273-82. doi.org/10.17113/ftb.61.03.23.7843
Figures can be graphs, line drawings, photographs or combinations of these. Check your target journal for whether multi-part images are common or whether they prefer to publish independent figures. The amount and style of figures required will influence your choice of journal. Some journals have detailed guidance on artwork, in addition to the basic guidance in the instructions for authors.
In photos of unfamiliar objects, either include a familiar object as some indication of size or insert a short rule and state its length (1 cm, 10 μm or another measurement relevant to your study.)
Check the journal’s requirements related to resolution (dots per inch, pixels per cm, etc.) and file format (eps, tiff, bmp, etc.). Especially with journals that are also published in print form, keep in mind the column width or page width to ensure that the figures can fit within the available space.
Example figure:
Figure 3. Delayed human NE transition is associated with a shorter cell cycle (A) Immunofluorescent staining for ZO1 on the surface of apical lumens of human (H9) and gorilla (G1) organoids revealing apical surface areas of individual neural progenitor cells at days 3, 5, 8, and 10. Perimeters of some individual cells of day 5 organoids are delineated in white. Scale bar, 10 μm.
Reproduced from Benito-Kwiecinski, Silvia et al. 2021 An early cell shape transition drives evolutionary expansion of the human forebrain. Cell, Volume 184, Issue 8, 2084 – 2102.e19 doi.org/10.1016/j.cell.2021.02.050
For a study where an initial sample is screened and eligible participants or materials are selected in some way, a study profile or flow chart is helpful. CONSORT has a template for clinical trials or interventional studies. PRISMA uses one for systematic reviews and meta-analyses. In both cases, start with the size of the total initial sample and give reasons for exclusion at each step.
Example flow chart for a meta-analysis reproduced from Li et al 2023. Efficacy of culturally adapted interventions for common mental disorders in people of Chinese descent: a systematic review and meta-analysis. Lancet Psychiatry, Volume 10, Issue 6, 426 – 440 doi.org/10.1016/S2215-0366(23)00118-9
Line graphs show how one variable changes with respect to another. They should be used for continuous data and the intervals along the x-axis should be consistent. More complex line graphs can be used, if necessary, but choose the simplest format possible that best communicates your findings.
When the data are not continuous (e.g. when showing the effect of different mass fractions of a compound on the product yield), use a histogram (bar chart).
Example figure: Fig. 4. Adalimumab yield and dsDNA quantity after precipitation and dissolution of adalimumab using different mass fractions of PEG4000
Reproduced from Vaskó et al. 2023 Development and comparison of alternative methods for the purification of adalimumab directly from harvested cell culture fluid. Food Technol Biotechnol. 2023;61(3):339-49. doi.org/10.17113/ftb.61.03.23.8094
Sometimes it can be helpful to summarise your findings in a diagram or model. Keep this simple, with as few lines as possible, and label clearly.
Example model reproduced from Buckton et al 2023 One Earth, Volume 6, Issue 7, 824 – 842 doi.org/10.1016/j.oneear.2023.06.006
The box and whisker plot is useful for displaying the distribution of a data set. It presents minimum value, first (lower) quartile, median (average value of the data set), third (upper) quartile and maximum value. It also helps to detect outliers, i.e. the data points outside the whiskers of the box plot, which do not follow the general trend.
Example figure: Fig 2. a-b. Amphora origins and contents. Box and whisker plots showing the percentage of each origin and content as part of each site assemblage.
Reproduced from Franconi et al. (2023). From Empire-wide integration to regional localization: A synthetic and quantitative study of heterogeneous amphora data in Roman Germania reveals centuries-long change in regional patterns of production and consumption. PLOS ONE 18(1): e0279382. doi.org/10.1371/journal.pone.0279382