Presentation on theme: "Presenting Statistical Data: Psychology, Misconceptions, Misperceptions and Other Illusions Philly CMG – May 18, 2012 Pete Warchol, MCT, MCITP, VCP, CEH."— Presentation transcript:
Presenting Statistical Data: Psychology, Misconceptions, Misperceptions and Other Illusions Philly CMG – May 18, 2012 Pete Warchol, MCT, MCITP, VCP, CEH
About the presenter Pete Warchol has been in the Information Technology industry for more than 20 years. He is a consultant and an IT trainer, plus a CMG Officer, for the Philadelphia regional group (http://regions.cmg.org/regions/phcmg/index.h tml).http://regions.cmg.org/regions/phcmg/index.h tml
Overview A discussion of how visual presentation scale, window of time, careful phrasing, sample size and implied comparisons can give an impression that the raw data does not exactly support. Examples from day to day life will be presented. The audience will be encouraged to start to think differently about the data and what questions they should be asking, especially when presented with a provocative statement or metric.
Concepts to be covered, in no particular order, and some examples will involve more than one of the following: visual presentation scale window of time careful phrasing sample size implied comparisons
Which is better or worse?
Now, which is better or worse?
WARNING: inflammatory example The next example has been chosen with malice of forethought, for the expressed purpose of driving home the point.
Have you ever heard either of these statements? The number of people paying into Social Security, versus the number of people collecting, has dropped from 42:1 to 3:1. or The number of people paying into Social Security, versus the number of people collecting, has dropped from 17:1 to 3:1. They are both completely true! However, …..
What do you think that graph looks like, for the past 70 years? Please form a mental image in you mind.
Was it any of these?
The raw data came from
Fifty percent of the people in the World are below average! Is that true? Do you think that it’s awful? Did you fall for the terminology trap? Will an “average” person fall for the terminology trap?
What the graph looks like.
Is average even meaningful? It depends! Take the following example of 100 instances: 1 instance of 1 98 instance of 10 1 instance of 1000 The average is 19.81, but the normalized average is 10.
Have you ever heard an advertisement that stated the following: A survey showed that 4 out of 5 dentists recommend sugarless gum, for their patients that chew gum. Have you ever critically analyzed that claim?
The questions that should come to mind. What did the 5 th dentist think? How many dentists did they ask? What about patients that don’t chew gum? How old is this survey? Who conducted the survey? Who paid for the survey? Did the dentists have an opinion about the specific product being advertised? So what, anyway?
The history behind 4 out of 5 dentists From: Activities/Articles/A0471-No-Gum-at-All-1-in-10http://www.bookofodds.com/Daily-Life- Activities/Articles/A0471-No-Gum-at-All-1-in-10 It comes from a 1976 survey About 1200 dentists were surveyed 85% answered that they’d recommend sugarless gum and most of the rest recommended no gum, at all. I’d like to find the full survey report.
Any questions or comments, before I present my conclusions? ???
Conclusions There are many ways that the presentation of data can give a misleading impression. Some people mislead by accident and some with deliberate intent. Failure to understand the detailed meaning of the data can lead to very poor decisions. Ask the pointed questions! Encourage the pointed questions!