Open Mind

Picking Cherries

November 12, 2006 · 3 Comments

There’s a website called co2science which regularly posts a “Temperature Record of the Week.” This week they’re featuring Beatrice, NE. From their website:

To bolster our claim that “There Has Been No Net Global Warming for the Past 70 Years,” each week we highlight the temperature record of one of the 1221 U.S. Historical Climatology Network (USHCN) stations from 1930-2000.

This issue’s temperature record of the week is from Beatrice, NE. During the period of most significant greenhouse gas buildup over the past century, i.e., 1930 and onward, Beatrice’s mean annual temperature has cooled by 1.07 degrees Fahrenheit. Not much global warming here!


Well … that settles it!

Of course you might wonder, “Is one single location truly representative of the entire globe?” Of course not! “Why choose the time interval 1930 to 2000?” Because that’s the time interval which gives them the best chance to make their claim look good.

Suppose I wanted to persuade Americans that temperatures aren’t rising — whether it’s true or not. What temperature record would I show? If I use the entire globe, it just won’t work, global average temperature has gone up. So, I’ll have to use a smaller area; after all, small samples (say, one meteorological station) tend to show much more variation than averages over large areas (say, Earth). I’ll have to pick a location in the U.S., because that’s my target audience — if I pick Marseilles, France, my American readers will wonder “why France?” but if I find a spot in the good old U.S.A., it’ll seem entirely natural. So, the logical choice is to pick a single meteorological station from the U.S. I’d better leave out Alaska, which is warming so fast that it threatens the lifestyle of the natives, and Hawaii is so distant and so ocean-locked it probably won’t “ring true” with my readers. So I’ll restrict myself to the 48 coterminous states of the U.S. Never mind that even if you used the entire area of the 48 coterminous states of the U.S., you still only have about 1.5% of the area of the planet. I’m not trying to get at the truth, I’m trying to support a given belief.

What time period should I choose? I have to make it seem as though it’s “up-to-date,” but I’d rather not actually be up to date, because temperatures have been higher over the last decade than ever recorded before. So, I’ll end at the year 2000, which seems to most people to be “current” but actually omits the time period (2000-2006) that most strongly contradicts my belief. I’d rather not start at the beginning of instrumental records, because then it’s just too difficult to find locations that’ll make my case look good. If you study the average temperature in the continental U.S., you’ll see that it rose in the early part of the century, peaking around 1930. Then it declined until about 1970, since then it’s been rising dramatically. Since it peaked around 1930, I’ll choose that as my starting time. In case anybody thinks 1930 is an unusual choice, I’ll call 1930-2000 the “period of most significant greenhouse gas buildup.” Too bad that when it comes to greenhouse gas buildup there’s nothing at all special about 1930 (look here and decide for yourself), but that doesn’t matter, I don’t expect my readers to check the facts anyway.

Clearly the best way to make the case is to choose single reporting stations from the continental U.S. and isolate the time from 1930 to 2000. What does the co2science website do? They choose single reporting stations from the continental U.S. and isolate the time from 1930 to 2000.

There’s a name for this approach to presenting information: cherry-picking. The idea is to display prominently the data that makes my case look good, omit whatever makes my case look bad. If you’re a lawyer in a criminal trial, this is what you’re supposed to do. Trial attorneys aren’t trying to discover the truth, they’re trying to persuade people to believe their side of the story. But if you want to be a scientist, this is not the way to do things. A scientist is not an “advocate” of a given opinion, but should be a discoverer of the truth. This requires presenting all the data, whether it supports a pet theory or not.

Consider that in the public debate over global warming, the denialists consistently use cherry-picking (and other subterfuges) to misrepresent the facts in order to make their case. Climate scientists consistently present all the data in order to get closer to the truth. When co2science shows you a temperature record, it’s a single station in the U.S. from 1930 to 2000. When James Hansen (NASA’s top climate scientist) shows you a temperature record, it’s from planet Earth and includes all the data we’ve got.

In case you’re wondering what happened in Beatrice, NE, before 1930 and after 2000, here’s a graph of all the available data from the global historical climate network:


During the period covered by the available data, Beatrice’s mean annual temperature has warmed by 0.33 degrees Celsius (0.59 degrees Fahrenheit). Not much global cooling here!

Categories: Global Warming

3 responses so far ↓

  • Peaseblossom // November 13, 2006 at 6:00 pm | Reply

    “Who is the bigger fool: the fool, or the fool that follows him?” I’m astouned that someone would actually use this method and try to pass it off as legitimate science; but what stuns me even more is that people fall for it.

    I’ll be working on that guest blog for you today, and it mentions “cherry picking” as well.

  • ladybarnard // November 13, 2006 at 7:31 pm | Reply

    We were actually taught this method of displaying statistical information in grad school. If you ever want to be truly scared by how easily information can be twisted, take “Visual Display of Quantitative Information, Introduction to Information Design” at University of Baltimore. One snippet we learned: bad information design directly led to the destruction of the space shuttle Challenger.

    The truth is, you can persuade anyone of almost anything if you’ve got pretty charts to back it up. Almost no one bothers to drill down into the actual numbers. I call it “USA Today” syndrome.

    Never forget the famous aphorism: “There are lies, damned lies, and statistics.”

  • Global Warming: Conclude yourself « Environmental Engineering News and Sharing // January 2, 2007 at 9:11 am | Reply

    [...] What do global warming denialist do to support their claim A certain global warming denialist website, co2science, regularly posts a “temperature record of the week.” I showed how they use clever cherry-picking to make their case look good in another post. But hey, two can play at that game; I too can post a “Temperature Record of the Week.” In fact, I’ll post two. [...]

Leave a Comment