GER1000 Tutorial 3: Sampling

Taking a sample of the population and estimate base on the data needed.

1. Accuracy:
Simple random sampling:
Each person got an equal chance of being selected

Systematic Sampling:
This is useful when you don't know the exact size of the population.
If a person refuses to talk, take it as a non-response.
Every survey should report on the survey rate.
Probability sampling plan definition:
Every unit have a known probability to be sampled

Stratified Sampling:
Calculation naturally falls into many subgroups e,g faculty of science, engineering.
Take an equal amount of sample from each subgroup


Cluster Sampling:
Randomly select a cluster and interview everyone in the cluster.
Chances of selection are the same

Multi-stage sampling
Divide the sampling frame into stage before repeating till we get a unit.
Chances of selection is the different

Volunteer Samples:
Getting people to answer, volunteer rather than finding people.

Difference between Stratified and Volunteer:
If we can calculate the response rate, its a volunteer sample

X large sample size does not mean no bias
Our frame should be bigger than our population



No comments:

Post a Comment