Monday, June 22, 2009

Climate Change and Information Design

A great entry on RealClimate, titled Curve Manipulation Lesson 2, shows the importance of proper charting techniques. I particularly like the postscript:

p.s. Beck appeared on German TV last Monday, after the “Swindle” film was shown, and he is announced to appear on the program “Report M√ľnchen” in the first channel of public German TV next Monday (18 June), to educate the viewers about another of his fantasy graphs, namely his CO2 curve. It promises to be a must-see for friends of the unintentionally farcical.

Edward Tufte refers to this as “graphical integrity” and it is one of the basics in his classic The Visual Display of Quantitative Information. I share ET’s passion for getting this stuff right because it matters and:

  • Done right, it can illuminate information and greatly improve communication by making complex ideas clear
  • Done wrong, it can obfuscate data and confuse or alienate people from otherwise good ideas

Good data display counts.

Designing useful questions

I received a survey recently that presumably was trying to get a sense of what the community thought:

Should we have an ordinance about land preservation?

(1) Yes
(2) Not sure
(3) No

The problem with yes/no questions like this is obvious – the topic is more complex than the question allows for, and the opposite of “no” isn’t always “yes.” Results from such a poll will be meaningless at best, which is fine for CNN's topic-of-the-day-ratings-boosting result, but not so good at understanding what a group actually might think about some complex topic. “Highest response” in this case is not the same as “majority opinion.”

A more useful question might have asked about different types of land preservation ordinances, listing them from extremely prescriptive and limiting to very suggestive and open to interpretation, as well as including a "no" choice and the ability to comment.

Polls and surveys are rarely useful in delivering the the sort of clarity needed to make good decisions or come to consensus. They are tools to help focus the questions, but not answer them.

Solving things ... or not

Human nature, and lots of us within our businesses, often rush to solve things – after all, that's what our clients pay us for, right? But sometimes, lots of additional dialog without a solution yields more value. The more complex an issue, the more unlikely it is that the immediate answer is the best one.

But listening can be difficult, especially when we really care about something or someone and believe that we know the answer. It can take time to really understand a different perspective, regardless of whether we agree with it or not. And many of us are impatient (I know I can be).

As we take more of our own conversations online, can we resist the lure of solving something before we really understand it?

Here we go

I've been blogging internally at my company for the last two years and a few people outside have asked if they could see the posts.

So here we (or I) go...