What is statistical error or bias? Lets say your cat is fat. You would like to measure the distance that your cat walks each day from your house to the house of your neighbourhood. Suppose that no obstacles exist and the cat walks this distance almost in a straight line.You do that in order to calculate the diet/calories of your cat. The first time your cat may have walked 5 meters, next time you measured 5.30 meters. The difference that exist between the next and previous observations is called statistical error. However, if the cat has died on the first few days of this experiment, or your pen has broken or some water has fallen on your notes, then your observations can have high statistical error. These events are random errors, they are not happen every day. Your glass of water will not broken every day or your cat cannot die everyday. However, if your measurement device has been stretched out from years of use, or each time you change this device then you may get systematically, each time different results. This is called systematic error. When additional observations are made , random error (in that case sampling error.) usually will decrease too. So, as more observations are made, their average number will deviate less (in average) by each one. Systematic errors are devious beings that are hiding very well most of times and it cannot be detected easily. Replications with cats located in another place or country (of same demographics such as age or race), or by adding multiple observers for example your neighbourhood, it can reduce Systematic error here. So statistics are useful or not? The cat do not care how much distance she did. Why must we care? More on the next video. Until that time, i am waiting your comments
Posted on: Sat, 03 Jan 2015 21:52:28 +0000
Trending Topics
Recently Viewed Topics
© 2015