answersLogoWhite

0

The Null Hypothesis. May be referred to as (H0). It is a statistical test and tricky in that this Null hypothesis can never be tested, so is tested in conjunction with some other hypothesis in such a way as the testing procedure, measurements made, will be done in a way as to show which hypothesis is more likely.

An example is given in Wikipedia regarding the tossing of a coin is worth viewing.

Null Hypothesis : we have a normal coin

Alternative Hypothesis one: we have a double headed coin

alternative Hypothesis two: Some unknown influence on an not normal coin

if a coin is tossed many times, say 100, and comes up heads each time the Null hypothesis is least likely of the three and REJECTED in favour of one of the other alternatives.

if the coin flips over and over and over easily but always lands heads up, most likely hypothesis two is correct and the Null Hypothesis REJECTED.

if the coin acts real strange by not going end over end and immediately falls very quickly down to land one particular side up hypothesis three is more likely and the Null Hypothesis REJECTED.

if the coin acts normally going end over end and lands about as many times heads as it does tails the Null Hypothesis is RETAINED as most likely.

The Null Hypothesis is either retained [kept as most likely true] or it is rejected in favour of the alternative. so it can be seen choosing which alternative to test against is always best and the simplest is generally the one to look at.

refer also Ocams Razor.

User Avatar

Wiki User

15y ago

What else can I help you with?

Related Questions