[mlthoughts2018]

Omitting sufficient detail to understand the scale and significance of a result is one of the most common types of manipulation in statistical data presentation. Not saying that this post was specifically trying to be manipulative, but it is very reasonable to presume the goal is manipulation and work back from there, since it is so, so common in published academic literature, newspaper infographics, etc.

[mikeash]

So, you’re not saying it’s manipulative, but it’s reasonable to assume it is? That seems contradictory.

[mlthoughts2018]

Why is it contradictory? I don’t have special knowledge to confirm if this was manipulative. But since manipulating the presentation of data is so common in academic papers, newspaper infographics, etc., it’s a reasonable prior belief. Generally being skeptical of data presentations is reasonable.