How to Lie With Charts

Rating:
Summary: Content fine, format hard to read
Review: A better title would be HOW TO AVOID BEING FOOLED BY CHARTS AND HOW TO PREPARE CLEAR INFORMATIVE ONES. This sounds pretty dull, but it isn't. The book is readable and interesting. The subject is also an important one, at least if you depend on information from charts. It is especially valuable if you regularly prepare charts for others. The title is sort of stolen from an even better book called HOW TO LIE WITH STATISTICS. That is an even more valuable book which teaches some of the same lessons as this one.

Rating:
Summary: Want to Learn How to Recognize a Deceptive Presentation?
Review: Are you tired of watching managements', employees' or politicians' deceptive presentation with graphs? Do you want to call their bluff? "How to Lie with Charts" is your secret weapon. This book along side with Darrell Huff's "How to Lie with Statistics" gives you all the amunition you need to cut through those presentations that create optical illusions.

The author explains all of the various charts available, their characteristics and how people alter their graphic works of art to influence the audience to buy into whatever the presentor wishes.

Not only does the author talk about the graphs but he explores the area of our subconcious and how this strongly influence our positive or negative perception of a chart.

The book goes into great detail and is quite humorous. The only cirticism that I have about this book is located in chapter 10. The author talks about the importance of color and how it influences the audience but he explains all of this in black and white. If you are going to encourge people to use color presentations and graphs, stop being such a tight wad and use color in your own book. Explaining tones, shades, etc., in fuzzy gray color doesn't do the job. Practice what you preach. Use color to explain color.

Rating:
Summary: Talk about yer irony.
Review: The writing is fairly engaging and the topic is covered fairly well, but the graphics are just awful!

In fact, the gray-scale images are so poorly reproduced that it detracts from the message the author promotes. And, in many instances, the graphs are so mottled that you have to guess as to what is being shown. Unless, of course, the book's poor image quality is part of the lesson. In which case it's a brillant ploy to get the point across.

Reasonable book - really lousy pictures!

Rating:
Summary: Very basic, very inconsistent
Review: This book could have value for some students, or anyone just beginning to create (or interpret) charts of quantitative information. Its friendly tone stresses the basics, like decent grammar, clear labeling, and thoughtful use of color. It also emphasizes the kinds of charts commonly created by PC desktop tools.

Lots of things get in the book's way, though. It dates from 1995, so lots of its advice about computer specifics (especially color) is dated. Although (fig. 10.4) he suggests underlining as a way to emphasize text, it was never a very good idea and has been given new meaning by conventions for hypertext links. Also, this edition is a black and white reprint of a book that seems originally to have been printed in color. That doesn't have to be a problem, but many of the original colors were replaced by annoying stipples or uneven screens, and one photo (11.2) makes no sense at all in black and white.

The worst problem is that the book often fails to follow its own good advice. Fig 3.14 uses cute 3D-ish cylinders as the bars in a bar chart, leaving the reader wondering just which part of the rounded end was meant to be the bar's end. Fig 7.8 is a supposedly good example tainted by the perspective distortion that Jones notes elsewhere. The "linked bar" of Fig. 7.10 is deceptive in that the bars themselves appear to cover some span of time, when they really represent instantaneous values.

Good advice, and there is a fair bit, often doesn't go far enough. P.105, for example, mentions log and log-log axes. Admittedly, the basic math is beyond this book's level (as shown by the blooper on p.140). Still, it wouldn't have been so hard to suggest using log scales for compound-interest or common growth curves, or to give a few examples where log-log scales would turn curves into straight lines with more meaning. Fig 7.17 shows two-sided bars (not usually recommended), without noting that they may be useful when the sum of the two sides has meaning. He discourages the use of "bubble" charts (p.128), even though that notation is common for chemists' phase diagrams or for placing different products along price and performance axes. He doesn't even mention the main value of "stacked" bars (fig 7.13), in showing the total of the stacked amounts. His "donut" chart (fig 2.10) is all but illegible, even though the same notation can be very useful for annotating one loop in many different ways.

Well, you get the idea. There is some good here, but lots that's not so good, including (p.200) reading your slides to your audience! I really can't recommend this one.

//wiredweird

Rating:
Summary: Excellent resource for algebra, precalc and computers
Review: This is a very funny book, but it's also quite informative. There are discussions of each kind of graph (or "chart") that you are likely to make, particularly if you use the spreadsheet software Excel. What types of graphs are appropriate for what types of data? When should you use a pie chart? How can you emphasize one piece of data in a chart, to make it stand out from others? This book answers these questions, and more. For algebra students learning to graph things on graph paper or on the computer, this may be interesting, or even more so for the teacher, who can use some of the funnier examples as a way to spice up the subject and keep students interested.

Besides discussions of the charts themselves, the author discusses how to write and display captions, how to put charts into slides, how to make an effective slide, how to change fonts and background colors to make your chart stand out, and more.

Reading this book will also help you to discern when other people have fooled with their charts to distort them. Local newspapers, news magazines, etc. are often guilty of playing with the scale of charts, stretching things, leaving labels off of axes, and so on - you'll be able to spot these manipulations better.

I teach a college freshman course in "Quantitative Applications Software" using MS Excel; I already have a lecture I usually call "How to lie with charts and graphs" and this book will help me add more details to that lecture, which teaches students that not every graph that CAN be made, SHOULD be made. With a good graph, you should always be able to start a sentence with "This graph shows that..." and complete it with some kind of comparison.

I have but one complaint about this book: it was clearly intended to be in a smaller format; each page of writing and illustrations takes up less than half the full-size page of the book. This could have been a trade paperback, and have cost less than it does as a larger book, without losing anything except 3" of empty margin all the way around. I plan to write to the publisher, telling them I really don't like that sort of inflation. However, you may find those margins handy for scribbling notes in; uses of this book are many, so you may need the space.