Statistics 4

[25362]

Are statistics indeed invaluable as evidence in support of conclusions ?

 

[BJC]

25362
Every drug you have ever taken, including any that may have save your life, was tested and then analyzed statistically.
Clinical testing and evaluation of drugs is probably one of the best gigs for PHD math types.

 

[Maui]

25362, is the employee who was responsible for 7 of your 10 annual safety "occurences" no longer working there?

Maui

 

[25362]

To all participants, thanks.

 

[arunmrao]

A PhD microbiology had drawn a curve with one point. Does he deserve the degree or needs to be pardoned ( as he does not know the criterion for plotting a curve)!!. This is a fact not a fiction.

 

[25362]

On the other hand, "significance", although not statistical in nature, can be attached to the total absence of occurrences.

For example, silo explosions of powdered Bisphenol-A during a March in Indiana, when motive air would be expected to be dry (with static electricity generation and no dissipation) when no such incidents happened in the humid environment of the Gulf Coast.

In this case "significance" refers to static charge generation as "probable" cause of dust explosions.

Would you like to comment ?

 

[IRstuff]

But, that's not what you asked:
"would a drop in that number to three be statistically significant?"

That phrase "statistically significant," is well-defined and understood in statistics.

Because, presumably, you have only a single data point, it may or may not be statistically significant, because, even if it is grossly out of the norm, it may well be a fluke; as alluded to by others, your plant's "Typhoid Mary" or "Martin" may have missed days for other reasons.

Even in your latest posting, significance can still be asked from a statistical perspective. You posit some number of explosions in a specific month, March, but presented no data for either location. Without the data, you cannot statistically demonstrate significance.

Likewise in your original posting and subsequent ones, you still have not presented any background to your numbers, hence, there can be no conclusion drawn from a statistical perspective.

TTFN