Please enable JavaScript.
Coggle requires JavaScript to display documents.
Sampling methods & sample size - Coggle Diagram
Sampling methods & sample size
Quantitative
Content Analysis
Existing Statistics (for meta analysis)
Surveys
Experiments
Why do sampling?
Cannot gather data from population
High cost & time
Not accessible
High bias --> Low confidance
probability value of error <0.05 significant (High level of confidance to generalise - reprodusability of our research)
p-value 0.02 means I am 98% confident that
Learn about Fuzzy Logic, bootstrapping (when small sampling size)
p-value cann't be applied for non-representative (non-random) sampling
terms
Sampling unit: smallest unit from which sample can be slected
Sampling frame:
Factors affect sample design
Research objective
Resources
KNowledge of target population
Degree of accuracy
Time frame
Research scope
Statistical analysis needs
Sampling Methods
Non-probablity
(not able to calculate sampling error)
Conveniance sampling
Snow-ball sampling
linear snowball sampling
Exponential non-discriminative snowball sampling
Exponential discriminative snowball sampling
Judgmental sampling/purposive sampling
Quota sampling: if three ethinic groups so each group select 10 people.
Probability Sampling
simple random sampling
Random number generator - random.org
Random table
Randombetween function excel
Systamatic random sampling
selection of every Kth element
k=(population size/sample size)
Stratified sampling
Findout Strata first then ramdomise within the strata
Cluster sampling
Multi-stage sampling
(a kind of complex samplie
Consider sampling weight - to compensate for under -over sampling
Inteer-cluster coorelation - need to calculate or consider as 0.02
Universal sampulation (no p-value calculation required)
Calculating sample size
Classical methds: rule of thump - lterature
Cochran formula is oK for prevance study only
Types of errors
Type error 1 - alpha
Type 2 error - beta
power = 1-beta
Most recomendation power accepted between 0.8-09
G power software for power based sample size calculation
Effect sizes
Small
Medium
Large
can get the details from lictature