FRIDAY SEPTEMBER 18, 2009

Cheating by Charting. An excerpt from Spear’s Practical Charting Techniques

There are few of us who, at at one time or another, have either exaggerated or shaded the truth by either bragging or playing down a story. What we say may not be an untruth, but we want to emphasize one fact to a certain party, and a different fact to another.

The same bragging or playing down can appear artfully in many types of charts. When and how do these distortions occur? They may be cleverly planned or happen unwittingly during the production of the visual.

Two distortions of the grid that occur most commonly:

1. The flexible grid: One of the easiest ways to give more or less movement to the trend of a curve is to expand or contract the horizontal or vertical axis of the chart. The chart below shows a correctly scaled trend and six ways the visual image changes by expanding or contracting the grid layout.

visual image charts_slide1

2. Skipping the grid: A familiar layout in reports and advertisements is seen here:

leftslide2

In order to dramatize the story, a little fudging is done with the time scale. It’s not noticeable at a glance that the time sequence is not uniform. It seems to be a neat, clean-cut, see-how-we’ve-grown story. Also, the dates being aligned parallel to the base line make the irregular date plotting less noticeable.

If you look at the chart below, it shows what the trend looks like when laid out with the correct grid spacing for each year. Amounts plotted for the given years are the same as in the chart above. Spread out this way is not as dramatic, but is the true story.

rightslide2

Topic: Book Reviews, Design, Strategy
Tags: charts, data, grid, trend, truth, visual image

COMMENTS (4)

Rob Hooft

^{September 20th, 2009

11:15 am}

Nancy,

A very often used tool in “lying” with graphics is changing the vertical scale. When a complete graph increases from 12 to 14 over a 5 year period, this does not look dramatic. A much more dramatic effect is reached when the vertical axis starts at 11. This effect can be used both ways, depending on which message you want to convey.

All About Presentations

^{September 22nd, 2009

10:29 pm}

Hi Nancy,

A very good observation and a matter of serious concern for presenters. Very often we take these things lightly and ‘tweak’ with our axis to highlight something which will never come to the fore if we start our axis from zero.

Should we always start our axis at zero?
I would say yes. It is always better not to distort the view of the audience to suit to one’s own liking.

However, if you do start with a non-zero axis you should inform your audience about it. I have seen umpteen charts, especially in annual reports and investor presentations, where axis starts with non-zero and the presenter never lets you know that the axis used is not starting at zero.

Look at Example 3b in the post below. I feel the change of axis is justified. However, what is missing is that the audience has not been intimated of the change.

http://www.allaboutpresentations.com/2009/09/14-charting-tips-tip-8-chart-axis.html

Regards,
Vivek

Thomas

^{October 2nd, 2009

3:44 pm}

I’m wondering why in the “correct grid” the space between 1990 – 2000 (10yrs) is 50% wider than 2000-2010 (also 10yrs).

Paresh shah

^{June 20th, 2010

1:46 am}

It has been a long tome since the post – so I am not sure whether this will be read.

In the graphs above, you have mentioned that one is correct and six wrong – the one which is correct has the following ratio : Vertical :horizontal ::3.5:4.5.

Is there a standard which is being referred to here – and what is the standard – charts are prepared for publication depending on the available size or the shape.

One would expect that in such situations there were would be standard set out to the followed by all.

SUBMIT A COMMENT

NAME (required)	COMMENT
MAIL (required, but will not be published)
WEBSITE