Lie factors

I've posted about my opinions over the "audience" of charts and graphics mattering. Take the above for example, direct from a school website. I wonder if the vast majority of parents would know what "Contextually Similar" meant and what the relevance of it is?

The y-axis is scaled from 86 to 98 - definitley non standard. Indeed, the 6% difference between 2012 and 2014 covers approximately 50% of the height of the graph - giving a lie factor of 8.33. I've reported before on lie factors bigger than 1.05 as being poor practice (see lie factors) (FYI - for this school with a standard cohort of 30-60 learners, this 6% rise corresponds to 3-4 learners)

Some Excel-fu on the chart, without changing much more than the axis error, to reduce the lie factor to 1.0:

More can be done though when we consider what we are asking parents to consider. Each temporal location is split into 4 series, so we are asking parents to compare 4 series across 3 time periods.

To track the progress of the school, they need to identify the right column (blue) and track it across the whole graph. Given that for most parents how the school is progressing in the most important data on the chart, lets rearrange the data:

I've also reduced the colours and chosen a darker shade for the school. Now I can track the school over 2012-2014.
I'm now wondering about that column "Contextually Similar" - without explaination (and even then) I wonder how useful that comparison is for parents - so let's remove it:

The primary comparison between School and Local Authority, I've overlaid and the National comparison I've kept to the right hand side.

Now I can see:
  • The 3 year progress of the school
  • How this relates to the Local Authority
  • Comparison to the National trend

comments powered by Disqus