This is another cautionary tale about data. It is also a plea to listen to what experts say about data despite headline news that show efforts by experts to manipulate data in various areas of life.
The diesel emissions scandal that has damaged the reputation of German automaker VW is a case in point. It took experts to develop the cheat system, and it took other experts to prove the company had circumvented emissions regulations for years.
And that does not include the frequent examples that show people believing in the correctness of their data, only to be proven wrong by some new discovery or method.
The blogster tries hard to avoid being duped, but has fallen for bad data many times - not counting the times it* has not realized having been suckered.
One aspect of events, and the data (or facts) they are based on, is whether they impact your life.
Which brings us to the exotic example of bird strikes. You may have heard of the birds that brought down a passenger aircraft in New York and made a hero of the pilot, Mr. Sullenberger.
So, birds striking a plane you are on can be life changing.
Collisions between birds and a military plane are generally considered part of the job and make few, if any, headlines.
So, it makes perfect sense that we don't know one interesting aspect of data on bird collisions with military planes.
Which is: military planes collide with birds much more frequently near the airfields than do civilian planes. The greater the distance of a fighter jet to its base, the fewer collisions are counted.
Why could that be?
There are many things that can come to mind, for example, whether military airfields are more frequently located in areas with more wildlife. Or maybe there is a correlation with the intensity of use of a base.
As the blogster learned, while this may influence the number of incidents, there was one other aspect that stood out: military pilots tended to report bird strikes close to the home base but ported fewer when they were far away.
As it turned out, protocol required pilots to find an airfield and have their bird inspected before taking to the air again. Which frequently meant the valiant fliers would not make it home for dinner.
Pilots and ground personnel felt protocols did not adequately handle bird strikes that caused no obvious serious damage to the plane.
They found a way around it, and in the process created "incorrect" data because collection was based on the two variables "bird strike" and "plane location", ignoring a third variable "extent of damage".
No comments:
Post a Comment