Design of tables for the presentation and communication of data in ecological and evolutionary biology
Data files
Aug 10, 2024 version files 258.08 KB
Abstract
Tables and charts have long been seen as effective ways to convey data. Much attention has been focused on improving charts, following ideas of human perception and brain function. Tables can also be viewed as two-dimensional representations of data, yet it is only fairly recently that we have begun to apply principles of design that aid the communication of information between the author and reader.
In this study, we collated guidelines for the design of data and statistical tables. These guidelines fall under three principles: aiding comparisons, reducing visual clutter, and increasing readability. We surveyed tables published in recent issues of 43 journals in the fields of ecology and evolutionary biology for their adherence to these three principles, as well as author guidelines on journal publisher websites.
We found that most of the over 1,000 tables we sampled had no heavy grid lines and little visual clutter. They were also easy to read, with clear headers and horizontal orientation. However, most tables did not aid the vertical comparison of numeric data.
We suggest that authors could improve their tables by the right-flush alignment of numeric columns typeset with a tabular font, clearly identify statistical significance, and use clear titles and captions. Journal publishers could easily implement these formatting guidelines when typesetting manuscripts.
README: Design of tables for the presentation and communication of data in ecological and evolutionary biology
Once we had established the above principles of table design, we assessed their use in issues of 43 widely read ecology and evolution journals (SI 2). Between January and July 2022, we reviewed the tables in the most recent issue published by these journals. For journals without issues (such as Annual Review of Ecology, Evolution, and Systematics, or Biological Conservation), we examined the tables in issues published in a single month or in the entire most recent volume if few papers were published in that journal on a monthly basis. We reviewed only articles in a traditionally typeset format and published as a PDF or in print. We did not examine the tables in online versions of articles.
Having identified all tables for review, we assessed whether these tables followed the above-described best practice principles for table design and, if not, we noted the way in which these tables failed to meet the outlined guidelines. We initially both reviewed the same 10 tables to ensure that we agreed in our assessment of whether these tables followed each of the principles. Having ensured agreement on how to classify tables, we proceeded to review all subsequent journals individually, while resolving any uncertainties collaboratively. These preliminary table evaluations also showed that assessing whether tables used long format or a tabular font was hard to evaluate objectively without knowing the data or the font used. Therefore, we did not systematically review the extent to which these two guidelines were adhered to.
Description of the data and file structure
In order to examine whether adherence to best-practice guidelines varies depending on the kind of information displayed in the table, we delineated three different types of tables. Statistical tables could range from correlation matrices to showing summary results of various statistical tests (e.g., mean, 25th and 75th percentiles, t/F/P values, degrees of freedom, etc.) to providing different kinds of information on models (e.g., probability at which a model is chosen for the given data, model predictions, model fitting times, model convergence rates, etc.). Text tables contained little or no numerical data. They often served to provide descriptions of measures/factors/analytical approaches used, or examples to demonstrate groupings/category labels, or summaries of different approaches/criticisms from the literature. Data tables either contained data in its unaggregated (raw) format or in a summarized form (e.g., summarizing the frequency of occurrences or categories of interest).
After reviewing issues in each of the 43 journals, we grouped descriptions of the ways in which tables failed to follow best-practice guidelines and tabulated how often the above principles were met. We also examined the formatting guidelines for tables provided by each journal. All data cleaning and analyses were carried out in the statistical software environment R 4.2.2 (R Development Core Team 2022).
The excel file contains three sheets.
Detailed results
This sheet contains the raw data collected on each table. Colunms provide identification information on each table, and then how that table met (or not) each guideline.
Big picture results
This sheet contains the cleaned and grouped data on each table. Colunms provide identification information on each table, and then how that table met (or not) each guideline.
Sheet 3
This sheet contains summary data on the tables in each journal.
Sharing/Access information
Links to other publicly accessible locations of the data:
NA
Data was derived from the following sources:
NA
Code/Software
NA
Methods
Once we had established the above principles of table design, we assessed their use in issues of 43 widely read ecology and evolution journals (SI 2). Between January and July 2022, we reviewed the tables in the most recent issue published by these journals. For journals without issues (such as Annual Review of Ecology, Evolution, and Systematics, or Biological Conservation), we examined the tables in issues published in a single month or in the entire most recent volume if few papers were published in that journal on a monthly basis. We reviewed only articles in a traditionally typeset format and published as a PDF or in print. We did not examine the tables in online versions of articles.
Having identified all tables for review, we assessed whether these tables followed the above-described best practice principles for table design and, if not, we noted the way in which these tables failed to meet the outlined guidelines. We initially both reviewed the same 10 tables to ensure that we agreed in our assessment of whether these tables followed each of the principles. Having ensured agreement on how to classify tables, we proceeded to review all subsequent journals individually, while resolving any uncertainties collaboratively. These preliminary table evaluations also showed that assessing whether tables used long format or a tabular font was hard to evaluate objectively without knowing the data or the font used. Therefore, we did not systematically review the extent to which these two guidelines were adhered to.