Skewed Statistics

Status
Not open for further replies.

bphgravity

Senior Member
Location
Florida
This past week, my division of the IAEI hosted the 76th annual Florida Chapter Meeting in Naples, Florida. One of the great aspects of these meetings is the quality face time one can have with some of the VIP's of our industry.

On one particular occasion Friday evening, I had a nice conversation with Mark Earley from the NFPA. One topic of discussion was regarding statistics in reports, proposal submittals, and other documents.

Mark told of an incident that occurred a few years back at an NFPA convention. At one of the banquet dinners, a poor soul jumped to his death from the fifth floor of the hotel right onto one of the guest tables. As it turns out, this was the second time an event similar to this happened to Mark's knowledge.

Mark's staff then decided to research the topic of suicide at hotels and found a striking statistic. 7 persons have jumped to their deaths from a balcony or atrium at a hotel on the same exact date the NFPA was holding a convention or meeting.

Based on this, a fairly reasonable person could conclude that there is a certain probability that a person will jump to their death from a hotel of greater than one story while the NFPA is also attending the hotel.

The point is that even though 2 completely and totally isolated events can take place at the same exact time and place, it does not mean placing some sort-of legislation or restriction on one will prevent the other from occurring. However, there are many instances where this type of philosophy has gone into NEC changes.

The bottom line is that statistics can be skewed to meet anyones interest and careful examination of the total facts is needed to make good decisions.
 

cowboyjwc

Moderator
Staff member
Location
Simi Valley, CA
It's kind of like the old joke about the man applying for the position of accountant at a large firm.

"So can you tell me how much $2,465,974.00 + $4,567,234.00 is?"

"How much would you like it to be?"
 

ceknight

Senior Member
Re: Skewed Statistics

bphgravity said:
...a fairly reasonable person could conclude that there is a certain probability that a person will jump to their death from a hotel of greater than one story while the NFPA is also attending the hotel.

The point is that even though 2 completely and totally isolated events can take place at the same exact time and place...

That same reasonable person could hypothesize that the events are not completely and totally isolated. If the correlation probability is high enough to get predictive power, heck, there could be several research grants in it.

In the meanwhile, perhaps NFPA could require hard hats at their conventions and offer suicide counseling for all of the hotel guests. If they're philosophically bent on legislating correlations, it's only fair. :)
 
Status
Not open for further replies.
Top