You are on page 1of 5

SAMPLING DESIGNS/TECHNIQUES

WHY SAMPLE?
In many fields of investigation, the researcher may use one of the two research designs illustrated
below.
Research Design I Research Design II
Population Sample

Descriptive Descriptive
Statistics Statistics

Inferential
Statistics

Parameter Statistics

The research design I requires the total units under investigation known as population.
The method of gathering the facts of interest on every unit of the population is called census. It
is well known that it is not always possible to get timely accurate and economic data by the use
of census. The method that is widely used nowadays is sampling.
Sampling is the process of selecting a part called sample from a given population with
ultimate goal of making generalization about unknown characteristics of the given population.
This is shown in research design II.

ADVANTAGES OF SAMPLING
1. Sampling enables the investigation of a large population.
When the population is too big, then it is almost impossible to collect data from all the
elements of the population.
2. Sampling reduces cost.
3. Sampling enables the completion of the study within a reasonable period of time.
4. Sampling avoids consuming all the sources of data.

SAMPLE SIZE DETERMINATION


There are so many ways of determining the sample size one of which is that of the
formula of Slovin (1960):
N
n=
1+ N e 2
Where:
n = a sample size
N = population size
e = desired margin of error (any percent value less than or equal to 10%)

Page |1
Example:
If the population consists of the College of Health Sciences students enrolled this
semester with a size of 1600, what could be a good sample size for a survey involving these
students?
Solution.
Given: N = 1600
e = 5% (a number within the range of 1% to 10%)
Required: n = sample size
Equation:
N 1600 1600 1600 1600
n= = = = = =320
1+ N e 1+[1600∗( 0.05 ) ] 1+(1600∗0.0025) 1+ 4 5
2 2

Conclusion:
A group of 320 students from the CHS constitutes the sample.

Always remember, however that the assumption of a normal distribution of the


population should be considered. When the normal approximation of the population is small or
poor, tis sample size formula does not apply.
Gay (1976) offers some minimum acceptable sizes depending on the type of research as
follows:
a. Descriptive research – 10% of the population. For smaller populations, a minimum of
20% may be required.
b. Correlational research – 30 subjects.
c. Expost facto or causal comparative research – 15 subjects per group.
d. Experimental research – 15 subjects per group. Some authorities believe that 30 per
group should be considered minimum.

SAMPLING TECHNIQUES
Once the sample size is determined and the list of population elements is available, the
next question to answer is "How could the sample elements be selected from the population
elements?"
The basic principle to remember in the process of selecting is "the sample elements
should truly represent the population elements". This means that characteristics of the sample
may or may not be taking all the characteristics of the population and that sample elements
should not contain any characteristics not found in the population elements.

1. Probability Sampling Techniques


There are techniques that allow every element of the population an equal chance of being
selected as a sample element. The selection may be done using:

I. Random Sampling is the method of selecting a sample size (n) from a universe (N)
such that each member of the population has an equal chance of being included in the

Page |2
sample and all possible combinations of size (n) have an equal chance of being
selected as the sample. There are several ways of drawing sample unit at random, it
can be done by:
a. Lottery Sampling or
b. Table of Random Numbers
c. Use of Calculators
a. Lottery Sampling. The lottery sampling method is usually carried out by assigning
numbers to each member of the population. For example. we may write down the names
of each member of the population on pieces of paper. These papers are then placed in a
box or container drum. The box or lottery drum must be shaken thoroughly to prevent
some pieces of paper from sinking at the bottom, where they will have less chances of
being drawn. From the box or lottery drum, the required number of sample units are
picked.
b. Table of Random Numbers. The use of the Table of Random Numbers is another
example of random sampling. Under this technique, the selection of each member of the
population is left adequately to chance, and every member of the population has an equal
chance of being chosen.

c. Use of Calculators. Some calculators have a key labelled RAN that gives random
numbers. The numbers that appear when this key is pressed have three decimal places. If
the population is less than a hundred, select the first two digits and disregard the decimal
point. The numbers are either one or two digits. Consider the following example:
Ran# Interpretation
0..185 18
0.284 28
0.678 67
0.726 72
0.410 41
0.014 1

2. Systematic Sampling. This method uses prior knowledge of the individuals comprising a
universe with the end in view to increasing precision and representation of samples.
When sample units are obtained by drawing every, say 4th or 7th or 10th item on a list, the
process of selecting the sample is called systematic sampling.
N
To get the Kth interval, we use . We usually get a number from 1 to K for a random
n
start. All other sample numbers are readily obtained by adding K to the previous number.

Example:
Suppose a sample of 75 students is to be chosen from the population of 325 students.
Identify the sample numbers by systematic sampling with a random start.

Solution:
N 325
Step 1: = =4.33 ≈ 4
n 75
Step 2: Pick a starting point using lottery method (random start).

Page |3
Step 3: add K to get the next element. (Example 3+4=7, 7+4=11, 11+4=15).

3. The Stratified Sampling. In this method the population is first divided into groups -
based homogeneity - in order to avoid the possibility of drawing samples whose
members come only from one stratum.
Stratified sampling is often used in polls of public opinion in order to secure
representative proportions of opinions coming from various classes of people.
Classifications may be based on districts, socio-economic status, sex, work, etc.
depending on the problem being studied.

The sample size per stratum is obtained using the formula:


population ¿× desired sample ¿ ¿
subpopulation ¿ ¿ ¿
¿

Example: A stratified sample of size n=500 is to be taken from a population size of N=4000,
which consists of three strata of size N1=2000, N2=1,200 and N3=800. If the allocation is to be
proportional, how large a sample must be taken from each stratum?
Solution:
2000 1200 800
n1= ×500=250 n2= ×500=150 n3= ×500=100
4000 4000 4000
n1 +n2 +n3 =250+150+100=500=desired sample ¿ ¿

4. Cluster Sampling. The cluster sample is sometimes referred to as an area sample


because it is frequently applied on a geographical basis. On this basis, districts or blocks
of a municipality or city are selected. These districts or blocks constitute the clusters.
Cluster sampling is useful in selecting the sample when blocks in a community or city are
occupied by heterogeneous groups. For example, if a community in Manila has lower,
middle-, and upper-income residents living side by side, we may use this community as a
source of a sample to study the different socio- economic groups in Manila. By
concentrating on this particular area, we can save more time, effort and money than if we
covered different communities throughout Manila.

A cluster is an intact group possessing a common characteristic. An example is a


population consisting of all the 800 nurses in 20 hospitals in a large city. Illustrate how
the desired sample size of 200 nurses could be selected given this population.
Solution:
Number of hospitals, N = 20
Average number of nurses per hospital, X = 40
Desired sample size, n = 200
n 200
Required clusters, y= = =5 clusters
X 40

Select the needed number of clusters. By using the Table of Random Numbers, select the 5
hospitals from the population list of 20 hospitals. Include all the members in the selected

Page |4
clusters. Since there is an average of 40 nurses per hospital and we shall only use 5 hospitals, our
sample size of 200 nurses is completed.

2. Non-Probability Sampling Techniques


1. Purposive Sampling. This is based on certain criteria laid down by the researcher.
People who satisfy the criteria are interviewed. A researcher might want to find out, for
example, the reaction of the banking community to a particular Central Bank circular.
Instead of interviewing the executives of all banks, he purposely can choose to interview
the key executives of only five biggest banks in the country if he believes that it is the
reaction of these big ones that counts anyway. Of course, the answers obtained through
this procedure are not representative of the entire banking system. Or a researcher may
want to find out whether the production of "burong talangka” conforms to the minimum
standards of health and safety. There are several small and medium-scale producers of
this product. However, to get a complete listing of the producers would be rather difficult,
What the researcher can do is to study and analyze only the two major producers of this
product.

2. Quota Sampling. This is a relatively quick and inexpensive method to operate. Each
interviewer is given definite instructions about the section of the public he is to question,
but the final choice of the actual persons is left to his own convenience or preference, and
is not predetermined by some carefully operated randomizing plan. Each interviewer then
proceeds to fill the prescribed quota. As the following example will show, this method
has its pitfalls. Suppose there is a survey to estimate what percent of the population of
Quezon City consider basketball as one of their favorite sports. One interviewer might
report that 100 percent of his quotas of 70 people are basketball fans. However, it may
later be found out that this interviewer reached his quota by going to the Araneta
Coliseum to enjoy watching his favorite team compete and at the same time interview
some thrilled viewers during the game.

3. Convenience Sampling. A researcher might want to find out the popularity of a radio
program. Since the researcher has a telephone, he might simply use it and "randomly"
pick his samples from the telephone directory. This method, of course, biased against
non-telephone users. Or a researcher might want to find out whether the production of
"bola-bola" or fish balls conforms to the minimum standards of health and safety. There
are hundreds of ambulant peddlers of this product. Thus, it is impossible for the
researcher to make a complete list, much less to interview all the producers and test all
their products. So, what the researcher can just do is to get samples of the product, say,
from the fish ball peddler near his school or near his residence.

Page |5

You might also like