Probability Sampling techniques all population members have a known probability

# Probability sampling techniques all population

This preview shows page 8 - 11 out of 16 pages.

2. ProbabilitySampling techniques: all population members have a known probability ofbeing selected in the sample. Most used to select large samples for conclusive research. (确定性研究)Why use probability sampling techniques?
Simple Random SamplingRandom sampling doesn’t happen randomly - each element in the population has a known and equal probability of selection.• Each possible sample of a given size (n) has a known and equal probability of being the sample actually selected.• You need a list for this• Appropriate for large populationWhy is simple random sampling (SRS) not used every single time?• Beautiful in theory• But difficult in execution if you don’t have list (how do you “randomly” select people ifyou are standing on the street? You cannot)• If you have several segments in your population, sample size for each segment will be proportional to their size in the population. Smaller segments might be missing Systematic Sampling（系统抽样）• Procedure – The sample is chosen by picking every i-th element in succession from the sampling frame通过从采样框中连续选取每个第i个元素来选择样本• When the ordering of the elements is related to the characteristic of interest, systematic sampling increases the representativeness of the sampleStratified Sampling (分层抽样)• ProcedureFirst divide target population into groups (strata)Picking elements from each group (stratum) using a probability based procedure• The strata should be mutually exclusive and collectively exhaustive (分层应该是相互排斥的并且是统一的) i.e., male and female are mutually exclusive and cover (almost)everyone• The elements should be homogeneous within a stratum, but heterogeneous across strata.元素是属于同质的阶层，但各阶层之间存在异质性。– i.e., on average, there are more similarities between male gamers (within stratum) than between male and female gamers (across stratum)• Used in cases where segments are of very different sizes and you want to ensure that all key segments are represented 用于分段非常不同的情况大小，你想确保所有的关键部分的代表Disparity in segment size: 98% are male, and only 2% female
ExampleTo conduct a survey of students’ usage of Monash library• Simple random sampling might yield insufficient number of students from smaller faculties (Buseco has nearly 10 times the students of Architecture)• Stratified sampling – each faculty is a stratum（阶层）and we randomly pick a numberof students from each faculty( need to define which stratum is and what is your sampling frame)Cluster Sampling• ProcedureDivide target population into groups (clusters)

#### You've reached the end of your free preview.

Want to read all 16 pages?

• Two '16
• representative, secondary data