Iona, de Dreu is a textbook example of how not to do diagrams. The two boxes of data you want to compare, i.e. the substance and placebo values of the same quantity are a long distance from each other. Instead they grouped data that have little relation to each other. The results are clearly distinct and the error bars don’t overlap, but they make it very hard for you to see that.
For all those still young enough to change bad habits, Paglieri offers good advice.
Sarnthein and Thornalley confirm you ought not to carbon-date marine samples. Not only is the reservoir effect unknown and varies geographically, it may also fluctuate by millennia over time.
Bohannon and Michel list an astonishingly long list of possibilities.
Here’s the link to this week’s complete list.