Friday, December 19, 2014

Awesome hall of shame devices - Interesting link

There's a cool article with ten hall of shame automotive devices that are/were supposed to do something or another to improve your vehicle's gas mileage, performance, and even ectoplasm removal (because I guess ghosts are slowing your car down؟). Popular Mechanic Gas Saving Gadgets covers it all.  My favorite is the magnets used to "straighten" fuel molecules to improve ignition efficiency. (If anything, the magnets would add more weight to your car, thus reducing fuel efficiency).

Saturday, December 06, 2014

"To good to be true" - Criticism of scientific studies grows

It is almost ironic, the other day I posted this article Reason Why I'm Skeptical of Skepticism which criticized over reliance on many study conclusions without actual supporting or valid data within the studies.  Now, just a few days later, there is a new published "study of studies" which reinforces the idea of being skeptical of scientific study conclusions, Excess Success for Psychology Articles in the Journal Science.  This study exposes that many studies in Psychology have issues, where the declared conclusions are simply "to good to be true" based on the strength of the data.  The inference being that there may be a general problem with all fields of science.
"Not every experiment is methodologically sound, and some experiments (even if methodologically sound) do not clarify the status of a theoretical idea. There is little reason to publish such experimental results, whether they are statistically significant or not. Unfortunately, in day-to-day scientific practice it is quite easy to interpret an unsuccessful outcome as being irrelevant to the theory or as being methodologically flawed and therefore not worth reporting."
In other words, data is cherry-picked in support of the theory rather than attempting to take contrary results into account.  This is basically throwing out the Scientific Method when it doesn't result in data this supports a theory.  In other cases, data collection is just too imprecise to form a suitable theory.  Kind of like garage-in-garbage-out.

I have a feeling a growing criticism of the current system is going to force changes into the process of study publishing and utilization.

Thursday, December 04, 2014

Old Perry Mason show becomes reality for Cops

What's up with all these Perry Mason style Grand Jury hearings being used to acquit police officers?  The latest case in New York is just nuts, where the accused officer was even allowed to testify, just like in the old Perry Mason show.  How are these Prosecutors allowing any defense at all during the hearing, let alone 2 hours of testimony by the accused without cross-examination (as in the Ferguson case)?  The cops are not being acquitted because of following the process. They are being acquitted because the Prosecutors are gaming the system to support the accused officers.  If a regular citizen is pulled up in front of a grand jury, they cannot present a case, they cannot challenge the presented evidence and they certainly don't get to testify for two hours without cross-examination.  Looks like it's time for other accused people to demand equal protection under the law, for equal treatment to these officers are getting during Grand Jury trails.