Download Sampling

Sampling
A famous sampling mistake
That’s Truman
They only asked rich, white
people with telephones
who’d they vote for. Sadly,
they published their mistake
What is a sample?
• A sample is any part of a population of individuals on whom
information is obtained: students, teachers, young learners,
etc.
• Selection of the sample of individuals who will participate in
the study is an important part of research.
• Sampling refers to the process of selecting these individuals
Samples and Populations
• A sample is the group on which information is obtained.
• The population is the larger group to which the researcher
hopes to apply the results.
• All university students studying English Language Teaching in
Turkey can be the population; the ones studying ELT in the
regions of Adana and Mersin can be a sample.
• Sometimes the sample and the population may be identical.
• Most populations are large, diverse and scattered, so it may
be difficult to obtain data from all. In that case, sampling is
needed.
• E.g.
• You are interested in the way English teachers are assessing
young learners in primary schools in Adana. There are 1,500
students in primary schools in that city. You can select 150
students in different schools as a sample for your study.*
• *The number generally depends on the methodology you will
use.
Defining the Population
• To define the population, we should answer the question,
“What am I exactly interested in?”, “What is the group to
whom I want to generalize the results of my study?”
• Some examples:
• All high school principals
• All fifth-grade classrooms in Mersin
• All language teachers teaching young learners
Target vs. Accessible Populations
• The actual population (target population) is rarely available. Then
the population to which a researcher is able to generalize is the
accessible population.
• E.g.
• Research Problem: The effects of computer-assisted instruction on
the reading achievement of 1st- and 2nd-graders in Turkey.
• Target population: All 1st- and 2nd-graders in Turkey
• Accessible population: All 1st- and 2nd-graders in Seyhan region of
Adana
• OR-- All 1st- and 2nd-graders in Celalettin Sayhan Primary School
• Sample: 10% of the 1st- and 2nd-graders in Seyhan region of Adana
• OR– 150 students attending 1st- and 2nd-graders in Celalettin Sayhan
Primary School
Random (Probability) vs. Nonrandom (Nonprobability) Sampling
• Random sampling (probability) means selecting the samples
without criteria (drawing 10 teachers out of 50 to interview)
• Nonrandom sampling (non-probability) means selecting the
samples based on a kind of criteria (the ones who have at
least 5 years of experience)
Scenario
• Hypo: Students with low self-esteem demonstrate lower
achievement in school subjects.
• Target population: all eighth-graders in Turkey
• Accessible population: all eighth-graders in Adana
• Feasible sample size: n= 200-250
Random Sampling Methods
1. Simple Random Sampling:
• The one in which each and every member of the population
has an equal and independent chance of being selected.
• If the sample is large, this is the best method.
• This can be done using a table of random numbers (can be
found in statistics books) or just drawing out the
names/numbers, etc.
Example on Scenario
• Identify all eight-graders in Adana (private and public schools).
Assign each student a number and select a sample of 200-250
students using a table of random numbers
• PS. Time-consuming to reach all schools
• 2. Stratified Random Sampling
• The process in which certain subgroups (strata) are selected
for the sample in the same proportion as they exist in the
population.
• E.g. If you want to compare students’ achievements regarding
their gender, you should ensure the proportion of males and
females is the same.
•
•
•
•
500 students (population)
200 males and 300 females
You want to use 20%
So you select 40 males and 60 females (20% from each group)
Example on Scenario
• Obtain data for all eighth-graders in Adana and determine the
proportion of each type (e.g. 80% public; 20% private)
• Public 80% of 200= 160
• Private 20% of 200= 40
• Randomly select students
• 3. Cluster Random Sampling
• Selection of groups of subjects, clusters, not individuals
• Used when it is not possible to select a sample of individuals (list of
all individuals not available, target populations is too big,
administrative reasons…)
• E.g. You want to see all elementary students’ attitudes towards
English. Not possible to get their names and use simple or stratified
random sampling. Then use some classes from selected schools
Example on Scenario
• Identify all private and public schools in Adana (having 8th
grade).
• Assign each school a number and select randomly 4 schools.
All 8th graders in these schools are your samples.
• Estimate of 2 classes per school x 30 students each x 4 schools
= 240 students
• 4. Two-Stage Random Sampling
• Combining individual and cluster random sampling
• E.g.
• first cluster sampling: select N number of classes from the
population
• Second individual sampling: select N number of students from
each class
Example on Scenario
• Randomly select 25 schools in Adana.
• Then randomly select 8 students from each
• 25 x 8 = 200
Nonrandom Sampling Methods
• 1. Systematic Sampling
• Selecting every Nth individual in the population.
• Get the names of all the population (alphabetically listed) and
select every nth number
• Be careful! if the names are not alphabetically listed (e.g.
listed according to the success level), your results may be
biased as you might not have any students who have
poor/high performance
Example on Scenario
• Identify the students in all schools
• Identify every 5th student if there are 1000 students in total
(250 students as sample)
• 2. Convenience Sampling
• Selecting individuals who are available.
• Generally, this sampling is not considered to represent a
population so is avoided. If this is a must, you should include
as much information about the sample as possible.
Example on Scenario
• Select all 8th graders in 4 schools to which you can access.
• Estimate of 2 classes in each school X 30 students X 4 = 240
• 3. Purposive Sampling
• Selecting samples based on researcher’s judgment
• Main disadvantage: researcher’s judgment may be wrong.
Example
• Select 8 classes from all schools on the basis of data you have.
• Be sure they are representative of all 8th graders
Choosing the method
Method
Best when
Simple random sampling
Whole population is available.
Stratified sampling (random within
target groups)
There are specific sub-groups to
investigate (eg. demographic
groupings).
Systematic sampling (every nth
person)
When a stream of representative
people are available (eg. in the
street).
When population groups are
Cluster sampling (all in limited groups) separated and access to all is difficult,
eg. in many distant cities.
Purposive sampling (based on intent)
You are studying particular groups
Convenience sampling (use who's
available)
You cannot proactively seek out
subjects.