7、Samplingwithunequalprobabilities.pptVIP

Chapter7. Sampling with unequal probabilities 1.Concept: If every unit in the population has an unequal probability to be chosen in the sample , we call it Sampling with unequal probabilities. The probabilities are often related with some auxiliary variables . such as units’ Size Measures , information highly related with survey variables 2.The condition of using it Sampling with unequal probabilities is often used under following circumstances: 1、To estimate the population total , but the the units’ sizes are six to one(相差悬殊）. 2、the units’ sizes are too large , then can not choose the basic relatively small units . 1.with replacement and without replacement According to whether the units replace to the population or not As to Sampling with unequal probabilities , replacement is often used 2 .PPS and PPZ According to either the probabilities are proportional to the units’ sizes or to the value of the auxiliary variable In practice, there are two ways to operating the Sampling with unequal probabilities . A) Hansen-Hurwitz method 累计总和法 Example: There are 8 streets ,we want to choose the streets according to the number of people it holds (10000) We should fix on a number by random from 1 to 45.1 , if 15 , so the street 3 is chosen ,because 15.1＞15. if 25,the street 6 is chosen (the random numbers from 24.3 to 33.1 are all mean the street 6 is chosen) The more people the street holds , the larger probability is. So , the higher the auxiliary variable is ,the larger the probability is . B) Lahiris method Example: There are 8 streets ,we want to choose the streets according to the number of people it holds (10000) Max{yi}=8.9，at first ,we fix on a number by random from 1to 8, if