Sampling (Trochim)

Simple

Stratified

how?

homogenous subgroup

sample

homogenous subgroup

sample

homogenous subgroup

sample

advantages

able to represent key subgroups in population, especially minority groups

more statistical precision (when subgroups are homogenous)

sampling fraction between strata

same

proportionate stratified random sample

different

disproportionate stratifed random sample

Cluster (AREA)

how?

1. divide population into clusters (usually geographically)

2. randomly sample clusters

3. measure all units withing sampled clusters

Systematic

how?

1. number units in population

2. decide on sample size (n)

3. calculate interval size (K = N/n)

4. randomly select integer between 1 and K

5. Take every Kth unit

advantages

easy to do

may be more precise than simple random sampling

Multistage

Convenience

Purposive

Modal Instance

sampling most frequent/typical case

Expert

sampling experts

Quota

proportional

nonproportional

Heterogeneity

sampling for diversity

Snowball

sampling hard to find populations

sampling frame

sample

standard deviation

1 stdev = 65%

2 stdev =95%

3 stdev = 99%

mean

sampling distribution

standard error

the spread of the averages around the average of the averages

sampling error

sample unit

measurement

response

statistic

examples

phone book

Subtopic

population parameter

generalization

sampling model

1. identify population you would like to generalize to

2. draw representative sample

3. conduct research on the sample

4. generalize results back to the population

problems

don't know what part of population you want to generalize to

not able to draw fair or representative sample

impossible to sample across all times you want to generalize to

proximal similarity model

gradient of similarity

dimension along which your study context can be related to other potential contexts to which you might wish to generalize

contexts

people

places

times

threats to

ways to improve

draw good sample

keep dropout rates low

use theory of proximal similarity effectively

provide data/descriptions

similarities

differences

do study in a variety of places