Small Data: Have we lost half the world's animals?
The most eye-catching statistic in the news last week was that the global animal populations have declined by 52% in the last 40 years, writes Anthony Reuben.
The figure comes from the Living Planet Report from the Zoological Society of London (ZSL) and WWF.
So there were half as many mammals, birds, reptiles, amphibians and fish (they only measure vertebrates) in 2010 as there were in 1970, according to the report.
This is one of those figures that makes you wonder how on earth they found out and indeed has sparked questions about both the statistical robustness of the figure and indeed whether having a single figure is relevant.
This is particularly the case because two years ago the same report said that the numbers had fallen by about 30%.
So what has changed since then? It's all in the weighting.
Saying things about a big group by finding out about a smaller group and extrapolating is never a perfect approach.
In the past, this report has looked at all the data available from species around the world and assumed that the trend seen in those species reflected the trend for all vertebrates worldwide.
Is that fair? This year's report followed data from about 3,038 of the estimated 62,839 vertebrate species, which is a pretty big sample.
But the sample is not random. There is much more research done into populations of birds and mammals than into reptiles, amphibians or fish.
The researchers have weighted the data to reflect what species actually exist and not just the ones that governments, academics or enthusiasts want to investigate.
This seems like a basically sensible approach, but there are still questions. For example, species that are in decline and a matter of concern may be more likely to be tracked than those that are not.
Also, species from less developed countries are more important to these figures than those in rich countries, because rich countries have often already lost areas such as forests and jungles where there is the greatest biodiversity.
So wildlife from the poorest countries get the greatest weighting in this research, but they are the ones for which by far the smallest amount of data is available.
Only 181 of the 3,038 species investigated came from low-income countries.
The researchers say that this is the first year they have had enough data to apply this weighting and that they will seek more figures for areas and species that are less represented to make future reports better than this one.
But while there are those who argue about whether the precise figure is accurate, there seem to be few who doubt the general trend.
Subscribe to the BBC News Magazine's email newsletter to get articles sent to your inbox.