Options

How to mislead with graphs -- example from pandemics

This is a chart making the rounds on Twitter that is supposedly showing that lockdowns are ineffective

There are so many issues with this chart don't know where to start. The citation they include is not the source of the chart but contains the data.

I think this awful chart came from this source, as this is the same guy that prompted the charts appearance

https://www.aier.org/article/the-bloodless-political-class-and-its-lack-of-empathy/

And the author likely grabbed it from this GitHub location, where I submitted an issue

https://github.com/tolex3/share/tree/master/20200727

The resolution is much higher in the original chart

https://github.com/tolex3/share/blob/master/20200727/lockdown_index_dead_per_M.jpg

Comments

  • 1.

    The first thing that caught my eye is that the chart has a negative y-axis, in spite of the fact that the units are in terms of deaths per population size -- which can never go negative. They did this to space the low death rates from the zero value so as to mislead the reader -- a common trick that Tofte taught everyone to be aware of.

    Comment Source:The first thing that caught my eye is that the chart has a negative y-axis, in spite of the fact that the units are in terms of deaths per population size -- which can never go negative. They did this to space the low death rates from the zero value so as to mislead the reader -- a common trick that [Tofte](https://www.edwardtufte.com/tufte/books_vdqi) taught everyone to be aware of.
  • 2.

    Oxford's "government response tracker index" tracks the stringency of government responses to COVID-19 across countries and time -- with 0 being least stringent and 100 corresponding to maximum stringency.

    The trend should be that the numbers to the left of the chart should be higher than the numbers to the right. Every sentient person knows this is the case with countries such as South Korea, New Zealand, Hong Kong which have a high index and few deaths, while countries such as Brazil have a relatively low index and higher death-rate. Yet this general trend does not show up on this chart

    Comment Source:Oxford's "government response tracker index" tracks the stringency of government responses to COVID-19 across countries and time -- with 0 being least stringent and 100 corresponding to maximum stringency. The trend should be that the numbers to the left of the chart should be higher than the numbers to the right. Every sentient person knows this is the case with countries such as South Korea, New Zealand, Hong Kong which have a high index and few deaths, while countries such as Brazil have a relatively low index and higher death-rate. Yet this general trend does not show up on this chart
  • 3.

    This person Jeffrey A Tucker that created the graph above is gaining a reputation of posting other misleading graphs

    https://twitter.com/CostaSamaras/status/1297014842604359680

    Comment Source:This person Jeffrey A Tucker that created the graph above is gaining a reputation of posting other misleading graphs https://twitter.com/CostaSamaras/status/1297014842604359680 ![](https://pbs.twimg.com/media/Ef_sHSHXkAAQ1Lh.jpg)
Sign In or Register to comment.