The Inspection Paradox 9/15/2015 • Allison Posner
Many are aware of the so-called friendship paradox of social networks, which was mentioned in one of the investigations in 1991 when social networks were offline. However, this paradox can be applied to modern online networks as well. This is actually a type of the famous inspection paradox.
The friendship paradox consists in the following: if we take any Facebook user and choose randomly any of his/her friends, we will find out with 80% possibility that the latter has more friends than the former.
In fact, there is nothing unusual. The essence of this paradox is that users who have more friends become a statistical sample more frequently. According to Stanford Large Network Dataset Collection an average Fb user out of 4000 users has 42 friends, but each friend out of these 42 has an average of 91 friends.
One may come across the inspection paradox quite often and it may even lead to delusions and mistakes. Here is a famous example: university students selected randomly are asked about the size of their group and according to their answers the arithmetic average is 56 people, though the university administration says there are only 31 students in each group on average. Surprisingly enough, but both the administration and students tell the truth. The thing is that people from larger groups have more chances to become a statistical sample just because there are more of them.
This mistake may seem quite trivial, but it is a reason for misunderstanding in many situations. The American professor of Mathematics Allen Downey provides an example of such a mistake. He observed the Red Line trains in Boston between 17:00 and 18:00 and arrived at a conclusion that the average interval between trains was 7.8 min. So, the average waiting time must have been about 3.9 min, but according to passengers the average waiting time was 4.4 min and the average interval between trains – 8.8 min. Thus, the difference was almost 15%.
The explanation is more than simple. During longer intervals, there are more people crowding on the platform and awaiting the next train, while during shorter intervals platforms and trains are less crowded. Obviously, there are more possibilities to select as a sample those people who are waiting during a longer period. The result is as follows: the majority of interviewed people complained about overcrowded trains and long interarrival time, whereas the company officials said that according to their data everything was within normal range.
And what do YOU think about it? Have you ever happened to come across this paradox?
How can overweight and obesity be prevented? Learn more about this major problem 10/21/2016 • Diane McKenzie
This educational video throws light on a problem which has become a real global epidemic of our time.
This aricle is for those who think that they saw everything and there is nothing to surprise them. But still nature has broad imagination. Welcome the most fabulous creatures in the world you don't know about!
Is coffee really harmful to your health? 9/27/2016 • Eli
I can't imagine my life without coffee, but there is always one question on my mind: how does it really affects my body and my brain?
"My name is Tim Doucette. I'm an amateur astronomer. And I'm also legally blind. Kind of an oxymoron, isn't?"
Science is the best way to keep kids entertained! By the way, can you explain all of these cool experiments?