SAMPLING

SAMPLING

EDUCATIONAL

RESEARCH

WHY SAMPLE?



 Not needed when can access entire

population.

 Must sample when not feasible to

access entire population.

Sample Should Be

Representative of the

Larger Population...

IMPORTANT TERMS



 Sample--subset of people selected

to study

 Population--larger group from

which sample comes to which we

will infer

STEPS IN SAMPLING



 ID ideal/preferred population

 Identify accessible population

 Use Principle of Sampling

Principle of Sampling



 If sample is representative of accessible

population, findings can be generalized

to population

 Generalizing from sample to population

involves risks

TYPES OF SAMPLING

PROCEDURES

 Probability

 Nonprobability

PROBABILITY SAMPLING



 Participants drawn by chance (random)

 Every member has known probability of

being chosen (1:10; 1:100)

 EPSEM--equal probability of selection

method

Probability Sampling: Types



 Simple Random Sampling

 Two-Stage Random Sampling

 Stratified Sampling

 Cluster Sampling

 Systematic Sampling

SIMPLE RANDOM SAMPLING



 All have equal and independent

chance of selection

Performing Random

Sampling

 Define and list ALL population

members

 Assign each a # from 0 to ????

 Group columns of digits according to

# needed

 Two digits for #s up to 99

 Three digits for #s up to 999

Performing Random Sampling

 Arbitrarily select a # in the random

number table

 If selected # corresponds to # of member

of identified population, member is in

sample

 go down # list

 repeat above selection method until

desired # of participants obtained

Two-Stage

Random Sampling

 Useful in large populations

 Select random clusters

 Select random individuals from random

clusters

STRATIFIED SAMPLING

 Information known about total

population prior to sampling

 Know population subgroups/strata

 Distinguish all elements in population

according to value on characteristic(s)

Performing Stratified

Sampling

 Identify subgroups/strata

 Select randomly a specific # of subjects

from each stratum

 Revisit random sampling

Cluster Sampling



 Unit chosen is NOT an individual, but a

group of individuals naturally together

 Constitute a cluster

 Are alike with respect to characteristics

relevant to the study

Performing Cluster

Sampling

 Chosen at random from population of

clusters

 All members of clusters chosen must be

included in sample

 Using clusters as individuals, follow

steps outlined in random sampling

Examples of Cluster Sampling



 Schools:clusters for sampling students

 Blocks: clusters for sampling residents

 Counties: clusters for sampling general

populations

 Businesses:clusters for sampling

employees

SYSTEMATIC SAMPLING



 Convenient to draw a random sample

when population elements are

arranged sequentially

Systematic Sampling



 Variation of simple random sampling

 Different: choices not independent

 Yields simple random sample in most,

but NOT all, sampling situations

 If sequence varies in regular, periodic

pattern, then will not have a random

sample---rarely occur

Performing Systematic

Sampling

 Determine size of sample

 Divide sample # into population

 Randomly select starting point

 Select every nth subject

 Need 5 people, have 45 in pop, select

every 9th person once starting point

chosen

Nonprobability Sampling



 Probability of selection of population

elements is NOT known.

 Participants NOT chosen by chance

 Cannot estimate likelihood selection

 More convenient and economical

NONPROBABILITY

SAMPLING TYPES

 Accidental Purposive

 Quota Snowball

 Convenience

 Cannot expect representative sample

using nonprobability methods

ACCIDENTAL SAMPLING





 Haphazard, availability, or convenience

 Interviewing/surveying first X number of

individuals encountered

 EXTREMELY weak, but fairly popular

 Most psych research is accidental

PURPOSIVE SAMPLING



 Subjects judged to be representative are

chosen from larger population

 Doesn’t produce representative sample

 Results may be misleading

PURPOSIVE SAMPLING

 Each sample selected for purpose,

usually because of unique position of

sample element

 Used in national elections (referred to as

bell-weather districts)

QUOTA SAMPLING



 Selection of typical cases from diverse

strata of a population

 Must know characteristics of entire

population to set right quotas

 Approximation of population with

respect to selected characteristics

SNOWBALL SAMPLING



 Useful for hard to reach or identify, but

interconnected populations

 May consider when cannot think of

another method

 Generalizations are very tentative

Performing Snowball

Sampling

 Identify one member of a population

 Speak with member

 Ask member to identify others

 Speak with those members

 Ask those members to identify more

 and so on and on and on....

Snowball Sampling

Examples

 drug dealers

 prostitutes

 practicing criminals

 gang leaders

 Alcoholics Anonymous members

CONVENIENCE SAMPLING

 When random is impossible

 Individuals who are available

 Likely to be biased

 Not representative of any population

 Should be avoided if possible

 Should be replicated

Sample Representativeness



 Sampling Goal: representativeness

 Larger the sample, more confidence we

have in representativeness of sample

 More homogeneous the population, the

more confidence of representativeness