Lies, Dam Lies, and Statistics
I used to do a lot of statistics; here are some choice stats funnies from years ago (from AllStat and elsewhere):
‘If I hadn’t believed it, I wouldn’t have seen it with my own eyes!’
‘First get your facts; and then you can distort them at your leisure’
‘There are 3 types of statistician: those who can count and those who can’t’.
The definition of a statistician is: ‘Someone who won’t accept that Columbus discovered America because he said he was looking for India in the trial plan.’
(c/o Patrick Royston)
There was once a group of Statisticians and a group of Epidemiologists riding together on a train to joint meetings. All the Epidemiologists had tickets, but the Statisticians only had one ticket between them. Inquisitive by nature, the Epidemiologists asked the Statisticians how they were going to get away with such a small sample of tickets when the conductor came through. The Statisticians said, ‘Easy. We have methods for dealing with that.’
Later, when the conductor came to punch tickets, all the Statisticians slipped quietly into the bathroom. When the conductor knocked on the door, the head Statistician slipped their one ticket under the door thoroughly fooling the layman conductor.
After the joint meetings were over, the Statisticians and the Epidemiologists again found themselves on the same train. Always quick to catch on, the Epidemiologists had purchased one ticket between them. The Statisticians (always on the cutting edge) had purchased NO tickets for the trip home. Confused, the Epidemiologists asked the Statisticians ‘We understand how your methods worked when you had one ticket, but how can you possibly get away with no tickets?’ ‘Easy,’ replied the Statisticians smugly, ‘we have different methods for dealing with that situation.’
Later, when the conductor was in the next car, all the Epidemiologists trotted off to the bathroom with their one ticket and all the Statisticians packed into the other bathroom. Shortly, the head Statistician crept over to where the Epidemiologists were hiding and knocked authoritatively on the door. As they had been instructed, the Epidemiologists slipped their one ticket under the door. The head Statistician took the Epidemiologists’ one and only ticket and returned triumphantly to the Statistician group. Of course, the Epidemiologists were subsequently discovered and publicly humiliated.
(c/o Brian Tiplady)
Two statisticians were on an evening flight from London to New York. Halfway across the Atlantic, a voice came over the intercom: ‘This is your Captain speaking. I have to tell you that the inner starboard engine of the aircraft has just failed. There is no need to worry, this plane will fly perfectly well on three engines, it just means that we will be about one hour late arriving in New York. Thank you.’
An hour later, the following announcement was made: ‘This is your Captain speaking. I have to tell you that the inner port engine of the aircraft has just caught fire. There is no need to worry, the engine is equipped with automatic fire-fighting equipment, and this plane will fly perfectly well on two engines, it just means that we will be about two hours late arriving in New York. Thank you.’
After another half-hour the aircraft gave a sudden jolt, after which came the following: ‘Ahem.. This is your Captain speaking. I have to tell you that the outer starboard engine of the aircraft has errrr dropped off…. There is no need to worry, this plane will ahem ahem fly perfectly well on one engine. It does means though that we will be about four hours late arriving in New York. Thank you.’
One statistician said to the other: ‘I hope that other engine keeps going, otherwise we shall be up here all night!’
– from the epidemio-l discussion list
Two guys are in a balloon, completely lost in a desert. Finally, they see a guy in the land. They ask : Where are we?? And the guy answers : in a balloon!!
The guys in the balloon must be epidemiologists because they formulate a correct question which does not address the problem.
The guy in the land must be a statistician because his answer, although correct, is totally useless.
Box and Whisky Plots
– a discussion thread initiated by Allan Reese:
A student’s report has pointed out to me an analysis technique that I shall in future undertake on every possible occasion and with gusto: "box and whisky plots".
Reply I: "The trouble with box and whisky plots is that they tend to make all distributions look skew or even multimodal."
Reply II: "Well, you just apply the Kolmogorov Smirnov test for normality, which consists of drinking vodka and looking at your data from time to time until it looks normal enough to you."
Reply III: "…. it certainly beats the tea-test …."
Reply IV: "This must be distinguished from the Kolmogorov Smirnov Two-Sample Test for Differences, which is viewed as being more robust than the Watney-Mann U (or Spew) Test. In the version of the K-S test I know, you need a quantity of vodka (amount determined from a "degrees of proof" procedure similar to power considerations) and your data. You look at your data.
A) *If* the relevant differences look statistically significant, drink the vodka, and look at them again. If differences still look significant, they are; if they don’t, get more vodka, and carry on with option B.
B) If the data do not look significant, drink the vodka as consolation."
Reply V: "…if what you describe is a two-stage procedure, then you need a washout period between periods A and B to eradicate ‘carryover’ or is it vodka-statistician interaction – I can never decide!"
Reply VI: "I believe that this is further documented in papers by Daniels, Jack (1984), Walker, Hiram (1986) and David, Mogan (1985)."
The First Significance Test
On AllStat someone enquired about who had done the first ever significance test . Allan Reese replied: And God divided the light from the darkness, and looked upon the light, and saw it was good with p<0.05. Later that same day; Many people have since suggested that this is also the first example of a Type I failure.