(9) Sampling Design

r

Davis, 2005

The Nature of Sampling

key terminology

Element

used to find data to form an analysis

Population

group that is under investigation

Sampling Units

nonoverlapping elements from a population

Sampling Frame

physical representation of all elements

Sample

a group from the population that is under investigation

Parameter/Statistic

summary descriptors of a given variable in the population/ " in the sample

Sampling Errors

errors in the study

Statistical and Sample Efficiency

Sampling Plan

formal specification(plan)

rationale

resource constraints

accuracy

destructive measurement

the sampling process

Step 1: Select the population

Step 2: Select sampling units

Step 3: Select a sampling frame

Step 4: Select a sample design

Step 5: Select size of sample

Step 6: Select a sampling plan

Step 7: Select the sample

The Feel Good Store (Project)

Sampling is important to our project because:

It can help to provide a number of conclusions in our study

can help find a less bias response, using the appropriate design

Some major considerations when choosing our design:

cost

time

An Introduction to the Philosophy of Sampling

goals

1. to acqaint you with the underlying philosophy of sampling

2. to show how and why samples can be used to estimate the parameters of a population of interest

probablity designs/ nonprobablility designs

Practical Considerations in Sampling

incidence and response rates

Incidence

% of the pop. that possesses the trait to be included in the study

Response rate

% of respondants contacted who participated

internationalization of the marketplace

Sample Designs

Probability designs

Simple Random

assign a number to each subject; select sample using the numbers

Systematic

natural ordering or order population

Multistage Random

use random ampling in each sampling stage where there are 2 stages

Stratified

a sample is selected from each stratum(non-overlapping groups)

Cluster

seperated into groups by proximity

Stratified Cluster

combo. of stratified and cluster

Repetitive

combining two or more of the designs above

Nonprobability designs

Judgment

selected by the researcher to rep. the population of interest

Convenience

convenience of the researcher is important

Snowball

samples are referred by the chosen population