Weissgerber et al. present a really nice examination of the problems with using bar and line graphs for data presentation in PLOS Biology. They discuss some of the downfalls of presenting data using summary statistics, as is the case with bar and line graphs. These problems include the fact that different data distributions can lead to the same graph, and therefore different conclusions; that it is difficult and even misleading presenting paired or non-independent data as bar graphs; and that summary statistics hide the underlying distributions behind the data leading to assumptions that the data are normally distributed and have no outliers. The final issue is particularly problematic for small sample sizes. They then go on to make recommendations to improve data visualisation: present a more complete picture of the data set; change journal policies to encourage better data visualisation methods; and improve statistics training to better include data visualisation. It’s a really interesting paper, and it has definitely made me rethink some of my figures.
This week has seen a few events of interest to quantitative ecologists. Take a look at the hashtags on Twitter to get an idea of what’s been happening at the BES Macroecology OpenData workshop (#besopensci), the Methods in Ecology and Evolution 5th anniversary symposium (#Methods5th) and the National Forum for Biological Recording conference (#NFBR15).
Courses The ALTER-Net Summer School, an interdisciplinary perspective on biodiversity and ecosystem services will be running in Peyresq, France from 2nd – 15th September.
If you have any suggestions of quantitative/computational ecology courses you would either like to attend or run, please fill out our survey here!
Jobs Post-doc at the University of Edinburgh (2.5 years) investigating the evolutionary ecology of a wild Soay sheep population. Research Associate position at UCL researching the biology of invasions by non-native avian species.