Thursday Highlights

9 Responses to Thursday Highlights

1. Boonton | February 26, 2009 at 11:45 |

   Why, I wonder, is that significant when (record?) polar ice advance is not? If weather is not climate … sorry but that’s not significant.

   What is important is the trend and the trend has been ice retreat. See, for example, http://www.nsidc.org/arcticseaicenews/faq.html#really_declining

   Yes, the data show that Arctic sea ice really is in a state of ongoing decline. The reason we know this is because satellites offer us a long-term record. As of September 2007, the September rate of sea ice decline since 1979 was approximately -10 percent per decade, or 72,000 square kilometers (28,000 square miles) per year. Although the 2008 sea ice minimum was slightly above the 2007 record, the rate of decline since 1979 increased to -11.7 percent per decade. September is the month that Arctic sea ice melts back to its lowest point, known as the annual minimum, and is an important indicator of overall ice conditions. However, sea ice in the Arctic is in decline in all months and the decline is greater and the rate faster than natural causes could account for.

   A record retreat of ice in one year then is simply an illustration of the underlying truth. It is like a report on how many Americans travel by plane that uses a picture or video of people boarding a 747. One could ask why not toss up a picture of a burning plane crash? Because while it’s clearly true that plane crashes happen (and of course the pic of the crash likewise happened), it doesn’t illustrate the report very well.

   When the datum outlier does become relevant is when it indicates a reversal of the trend or some very important change in what you are looking at. A report on air travel in the US might, if it was very comprehensive, include a 9/11 picture since that one event did have a major impact on air travel that we still feel today.

   So the question is assuming you have an example of a record polar ice advance is there a viable case that it meets the criteria of being cited as a datum outlier or not?

2. Mark | February 26, 2009 at 12:15 |

   Boonton,
   Over the last two (three) years, global mean tropospheric temperature has fallen. 9/11 is an outlier … but there is a identifiable cause. When ice advances or temperatures fall that might be an outlier. But no cause is cited … which is problematic then for your picture.

   Confirmation bias is another explanation for why retreat cited but not advance is noted.

3. Boonton | February 26, 2009 at 12:48 |

   I’ll tried to be subtle about it but now I’ll be more point blank, where is your hypothetical ‘record ice advance’?

   Now you’ve shifted to a different set of data (global mean tropospheric temps). I’ll point out that this has been reported on and also note that it does not contradict global warming theory. Here again we have you hyping one or two datapoints yet you’ll insist that global warming advocates produce thousand datapoint series to support their case.

4. Mark | February 26, 2009 at 12:52 |

   Boonton,
   No, I’m just pointing out that if hyping small-set data points on one side is meaningless so is hyping small-set data points on the other.

5. Mark | February 26, 2009 at 12:54 |

   Boonton,
   I should add I’m not in particular a denier of climate trends, just the epistemological certainty (or even reasonable probability) for human origins of the same.

6. Boonton | February 26, 2009 at 13:44 |

   Ahhh but hyping a small data set is ok if it is being done to illustrate a trend supported by a much larger dataset. If the trend is for ice melt in the artic then it is acceptable to use a one time dramatic ice retreat to illustrate that. The error comes from confusing the illustration with support. This can lead to the error that concludes a dramatic data point in the opposite direction is either newsworthy or, even worse, refutes the prevailing theory.

7. Mark | February 26, 2009 at 14:36 |

   Boonton,
   For your next spin assignment, you get to explain why citing less and more Atlantic hurricanes are both OK to be used as evidence of a warming trend.

   The error comes from confusing the illustration with support. This can lead to the error that concludes a dramatic data point in the opposite direction is either newsworthy or, even worse, refutes the prevailing theory.

   Exactly why the illustration is dangerous and should be avoided … which was my point.

8. Boonton | February 26, 2009 at 22:07 |

   Actually your point was that ‘record ice advances’ should be cited as heavily as ‘record ice retreats’ (such advances being totally hypothetical since there doesn’t seem to be any recent record advances).

   There’s nothing wrong with illustration. If you are doing a story on foreclosures it’s fine to pick out a family that is fighting to keep their home or just lost it. If you’re doing a story on diets use some chick who lost 100 lbs on Atkins. But don’t *reason* with illustrations. That’s simply my point.

9. Mark | February 26, 2009 at 22:51 |

   Boonton,

   There’s nothing wrong with illustration. [...] But don’t *reason* with illustrations. That’s simply my point.

   I really really really despise the use of illustrations especially when politicians trot out some heart wrenching narrative or even (more embarrassingly) drag to some event a family or person who’s story, oh so poignantly, illustrates why their particular (usually stupid) policy is required.