How not to read a graph

In two recent posts, Craig Kelly MP draws conclusions from a graph – and completely gets it wrong each time!

Global cooling graph (?)

On April 5th Craig Kelly tweeted this:

Missing the obvious

There are several reasons why this post is ignoring realities.

1. The overall trend on the graph is clearly upwards.
2. There are many short term rises and falls.

Climate deniers have claimed a cooling trend on other occasions, but it soon becomes clear that this was temporary. The graph shows that temporary fast rising and falling periods can last up to 5 years or more, so one cool year is too soon to claim a change in the trend.

3. We need to compare full years

It is surely obvious that temperature varies through the year (summer and winter), and we need to compare whole years with whole years! It is too soon in March 2021 to be talking of the annual temperature. When we compare whole years, 2020 was the equal hottest on record.

4. Cherry-picking is misleading

If we carefully select the two years we compare, we can find periods where the temperature appears to have fallen, or risen sharply. March 2021 may be cooler than many previous months, but most months in the last few years are hotter than previous months – because the trend is rising!

Mr Kelly has either misunderstood the graph or he doesn’t care that the graph proves him wrong.

Coronavirus graph

On April 3rd Craig Kelly posted this graph:

He added the comment that “the decline in deaths from the first wave (when there was no vaccine) virtually mirrors the decline in deaths from the second wave”. He wanted to question whether vaccination was making any discernible difference.

Doing the comparison

Just from the graph, the rates of decline don’t look the same. The second wave seems to drop faster.

So I cut and pasted the two graphs to overlay them, producing the graph at the top. Clearly the rates of fall don’t mirror each other at all!

So then I went to the actual data used by Mr Kelly and measured the speed of decline in deaths. I found that the second wave declined 50% to 75% faster (depending on which part of the graph I measured).

This is certainly nothing like the same as stated by Mr Kelly!

I haven’t tried to check whether this improvement is attributable to the vaccine or other causes, but it is clear that Mr Kelly has just made up his conclusion from a quick (and not very accurate) look at the graph.

How NOT to read a graph!

Mr Kelly has shown us how NOT to read a graph.

We need to see the context of the graph and what it is actually telling us. We need to read the data points accurately and understand them.

If not we will likely perpetrate misinformation and wrong conclusions.

Leave a Reply

Your email address will not be published. Required fields are marked *