Please enable JavaScript.
Coggle requires JavaScript to display documents.
Sampling (Sampling Process for Survey (Choose a sampling approach (2.3…
Sampling
Sampling Process for Survey
Define population
Target pop is determined by research topic
Some studies more target/ require focus on specific segments
Choose a sampling approach
sampling frame = list of customers to sample from
ask urself if have comprehensive lists of customers to contact/ if X sampling frame -- set of directions on how :silhouette: can be found
2.2 Specify survey method
if got list, draw names from there: telephone, mail
if set of directions: mall intercept, visit ppl at house
2.3 Pick sampling techniques
Non-probability ST
Probability unknown
base on judgment
small samples for exploratory research
Convenience Sampling
obtain a sample of conv elements
respondents selected- right place right time.
Judgemental Sampling
form of conv sampling, pop elem based on judgement of researcher
Etc, test mkt, expert witnesses in court
Quota Sampling
2-stage restricted judgemental sampling
1st stage: identify segment in pop
2nd stage: sample ele selected based on conv / judgement
Snowball Sampling
initial grp of respondents selected - convenience
After interview, respondent ask to identify others who belong to target pop of interest
subsequent respondent selected base on referrals
use to study illicit behav (illegal immigration/ drug use)
Probability ST
all pop member known probability
large samples for conclusive research
Simple Random Sampling
element in pop known/= probability of selection
need a list
large popu
Why not use everytime?
beautiful in theory
diff in execution X list
several segment in pop, smaller seg may missing
Systematic Sampling
Procedure: sample chosen by picking every i-th element in succession
whatever criterion you use to order elements= sample representation
Stratified Sampling
Procedure: divide target pop into grps, pick elements from each grps using prob-base procedure
Strata=
mutually exclusive/collectively exhaustive
elements should be
homogeneous within
a stratum but
heterogeneous across
strata
Cluster Sampling
Procedure: divide pop into grps
each grp by itself represent whole pop
mutually exclusive/collectively exhaustive
elements- heterogeneous within cluster but homogeneous across cluster
used in cases where pop is very dispersed/ data collection cost is high
Is sample generated "representative of target population?"
How to tell if miss segment?
knowledge of company/industry
secondary data