« Detached in time and space | Main | Good and simply better »



I couldn't agree more about the Faux accuracy!

Two other thoughts. The business about putting altitude or depth on the y-axis is common practice in many physical sciences fields like oceanography (see example at
http://www.windows2universe.org/earth/Water/images/temperature_depth_jpg_image.html), limnology and atmospheric science. I always find these charts hard to think about because they go against the standard practice (I'm an ecologist), but I suppose if you spend your days looking at them they start to make sense. The line joining a series of solutions reminds me of an "efficiency frontier" in economics, where you are trying to figure out how many different possible solutions compare. Good solutions are on the boundary of the cost-output space. This might be relevant if the report was arguing for an investment of say, 10^8 dollars in the Aurora design. Then the performance of Boeing's best alternative isn't relevant - only the performance of off the shelf solutions.

Laust Lund Kristensen

As you accurately pointed out, only the data in the 20 - 30 km range are viable alternatives to the hybrid aircraft. Since the closest datapoint in that range is a factor of 3 - 4 higher than the hybrid aircraft (we are looking at a log scale), surely error margins are not that big. So is faux accuracy really an issue here?


Laust: If the error bars are not plotted, then we don't know how big they are. Speculating on large-scale projects with new technology not yet implemented is a recipe for very wide prediction intervals. If they aren't large, they can't be believed. For example, Boston's Big Dig cost $15 billion with the original projection of $4 billion.
And remember, if Boeing gets its hands on this chart, I'm sure they won't feel fairly represented by that one point.

Aminpractice: I don't quite get how the line could be an "efficiency frontier" as I imagine that altitude is a driver of cost; there is no output measurement at all on this chart.

Laust Lund Kristensen

Kaiser: Thank you for taking the time to respond.

I agree with you that in principle all data should include some indication of error. But in some cases singificant digits is quite adequate as a measure of uncertainty. Here the datapoint for Hybrid Aircraft is at around 100 MUSD / year and the datapoint thethered pipe is at around 400 MUSD / year (the two closest datapoints in the range in question), so by the priciple of significant digits there is no overlap.

If you truly believe that the cost estimates can fluctuate by more than 100% then instead of demanding error bars, you should be demanding a clear overview of the model and its assumptions!

This comes back to your point about loss aversion as well. Since significant digits in this case likely understates the datas margin for error and this does not disturb the analysis of those data, then why should we demand error bars?

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.


Post a comment

Your Information

(Name is required. Email address will not be displayed with the comment.)


Link to Principal Analytics Prep

See our curriculum, instructors. Apply.
Marketing analytics and data visualization expert. Author and Speaker. Currently at Columbia. See my full bio.

Book Blog

Link to junkcharts

Graphics design by Amanda Lee

The Read

Good Books

Keep in Touch

follow me on Twitter