You are on page 1of 16

CHAPTER 2

PLANNING AND CONDUCTING SURVEYS

Chapter Objectives:
At the end of the lesson, students should:
1. Define and differentiate survey;
2. Discuss some potential pitfalls of a survey;
3. Explain the importance of survey in research work.
4. Enumerate and discuss the steps on how to plan and conduct surveys;
5. Define, identify and differentiate the different sampling methods/techniques;
6. Solve sampling data problems using sample size formula;
7. Solve sampling technique problems using systematic sampling formula;
8. Identify and discuss the sources of bias in surveys.

Introduction
Once we decide to get information from the sample not to conduct a census, the question that arises is
how to get a sample. How do we decide who receives the questionnaire or which item get tested? Does
it matter how to select a sample? As you will see, different situations warrant different methods of
sampling to address specific circumstances.

LESSON 2 – SAMPLING TECHNIQUES


Lesson Proper
In the collection of data, the number of sample and the nature of sample to be chosen are very critical
for the study to have reliable result. In this lesson, we will discuss the different ways on how to choose
samples.

Let us first define some terminologies that we will encounter in this section.
A population refers to the entire group that is under study or investigation.
A sample is a subset taken from a population, either by random or nonrandom sampling techniques. A
sample is a representation of the population where one hopes to draw valid conclusions from about the
population.

1. Probability Sampling
This sampling technique is also called the Simple Random Sampling. In this technique, the samples
are randomly picked and therefore the selection of sample is without any bias. Each member of the
population has an equal chance of being picked as part of the sample. A good example of this sampling
are the lottery and raffle.

Random sampling is a selection of n elements derived from a population N, which is the subject
of the investigation or experiment, where each sample point has an equal chance of being selected
using the appropriate sampling technique.

Random samples are selected by using chance methods or random numbers. One such method is
to number each subject in the population. Then place numbered cards in a bowl, mix them thoroughly,
and select as many cards as needed. The subjects whose numbers are selected constitute the sample.
Since it is difficult to mix the cards thoroughly, there is a chance of obtaining a biased sample. For this
reason, statisticians use another method of obtaining numbers. They generate random numbers with
a computer or calculator. Before the invention of computers, random numbers were obtained from
tables.

Some two-digit random numbers are shown in Table 1.1. To select a random sample of, say, 15
subjects out of 85 subjects, it is necessary to number each subject from 01 to 85. Then select a starting
number by closing your eyes and placing your finger on a number in the table. (Although this may
sound somewhat unusual, it enables us to find a starting number at random.) in this case suppose your
finger landed on the number 12 in the second column. (It is the sixth number down from the top.)
Then proceed downward until you have selected 15 different numbers between 01 and 85. When you
reach the bottom of the column, go to the top of the next column. If you select a number greater
than 85 or the number 00 or a duplicate number, just omit it. In our example, we will use the subjects
numbered 12, 27, 75, 62, 57, 13, 31, 06, 16, 49, 46, 71, 53, 41, and 02.

Table 1.1 Random Numbers


79 41 71 93 60 35 04 67 96 04 79 10 86
26 52 53 13 43 50 92 09 87 21 83 75 17
18 13 41 30 56 20 37 74 49 56 45 46 83
19 82 02 69 34 27 77 34 24 93 16 77 00
14 57 44 30 93 76 32 13 55 29 49 30 77
29 12 18 50 06 33 15 79 50 28 50 45 45
01 27 92 67 93 31 97 55 29 21 64 27 29
55 75 65 68 65 73 07 95 66 43 43 92 16
84 95 95 96 62 30 91 64 74 83 47 89 71
62 62 21 37 82 62 19 44 08 64 34 50 11
66 57 28 69 13 99 74 31 58 19 47 66 89
48 13 69 97 29 01 75 58 05 40 40 18 29
94 31 73 19 75 76 33 18 05 53 04 51 41
00 06 53 98 01 55 08 38 49 42 10 44 38
46 16 44 27 80 15 28 01 64 27 89 03 27
77 49 85 95 62 93 25 39 63 74 54 82 85
81 96 43 27 39 53 85 61 12 90 67 96 02
40 46 15 73 23 75 96 68 13 99 49 64 11

2. Restricted Random Sampling


This is often times used when the population to be considered is too large. There are two types
of restricted random sampling summarized as follows:
a. Systematic Sampling. The selection of sample is done by picking every k th element of the
population. The kth element in the population is obtained using the given formula below:

kth =
𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑠𝑖𝑧𝑒 𝑁
=
𝑆𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒 𝑛

Example: A researcher wants to conduct a study in a university with 10,000 students with
90% precision. If he used systematic sampling, find the k th member he must include
in his sample.

Given: N = 10,000; n = 99 (obtained from the previous problem)

Solution: kth =
𝑁 (𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑠𝑖𝑧𝑒) 10,000
= = 101
𝑛 (𝑠𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒) 99

Thus, every 101st member of the population was picked.

Systematic sampling – sampling technique in which members of the population are listed and
samples are selected in intervals called sample intervals. In this technique, every nth element
from the list is selected from a randomly selected starting point. For example, if we wish to
draw a sample of 2,000 from a population of 6,000, we can select every 3rd person in the list.
In practice, numbers between 1 and 30 will be randomly selected to act as the starting point.

Situational Example:
Suppose a quality control engineer is interested in checking the quality of items coming
off the automated production line producing 20 items per minute. Based on the inspected
items, he will decide whether the automated procedure is working as per specification.
Random sampling can become confusing and time-consuming, so he selects a random number
from the numbers between 1 and 20, say 15. Then he instructs the machine supervisor to
inspect the 15th item at the beginning of the day and then every 20th item coming off the
production line, inspect item 15, 35, 55, 75, 95, …. This method of sampling population is
known as systematic sampling.

In systematic sampling, the frame is divided into consecutive segments, a random starting
point is selected from the first segment, and then a sample is taken from the same point in each
segment (Figure a.1). This method is popular among biologist, foresters, environmentalist, and
marine scientist. It is also commonly used to sample people line up to attend events.

Figure a.1

Researchers
obtain systematic samples by numbering each subject of the population and then selecting
every kth subject. For example, suppose there were 2000 subjects in the population and a
sample of 50 subjects were needed. Since 2000 ÷ 50 = 40, then k = 40, and every 40th subject
would be selected; however, the first subject (numbered between 1 and 40) would be selected
at random. Suppose subject 12 were the first subject selected; then the sample would consist
of the subjects whose numbers were 12, 52, 92, etc., until 50 subjects were obtained. When
using systematic sampling, you must be careful about how the subjects in the population are
numbered. If subjects were arranged in a manner such as wife, husband, wife, husband, and
every 40th subject were selected, the sample would consist of all husbands. Numbering is not
always necessary. For example, a researcher may select every tenth item from an assembly line
to test for defects.

In systematic sampling, it is assumed that the elements of the population are arranged in some
natural sequential order. Then we select a (random) starting point and select every kth element
for our sample.

Example:
People lining up to buy rock concert tickets are “in order.” To generate a systematic sample of
these people (and ask questions regarding topics such as age, smoking habits, income level,
etc.), we could include every fifth person in line. The “starting” person is selected at random
from the first five.

The advantage of a systematic sample is that it is easy to get. However, there are dangers in
using systematic sampling. When the population is repetitive or cyclic in nature, systematic
sampling should not be used.

Example:
Consider a fabric mill that produces dress material. Suppose the loom that produces the
material makes a mistake every 6th meter, but we check only every 5th meter with an
automated electronic scanner. In this case, a random starting point may or may not result in
detection of fabric flaws before a large amount of fabric is produced.

b. Stratified Sampling. The population is divided into strata (groups) based on their homogeneity
or commonalities. The steps in doing the stratified sampling are as follows:
(1) Determine the distribution of the population in each stratum;
(2) Find the percentage of each stratum from the population;
Multiply the percentage of each stratum by the sample size (n).

Stratified random sampling – sampling procedure wherein the members of the population
are grouped based on their homogeneity. This technique is used when there are a number
of distinct subgroups in the population, within each of which is required that there is full
representation. The sample is constructed by classifying the population into subpopulations
or strata, based on some characteristics of the population such as age, gender, or socio-
economic status. The selection of elements is then made separately from within each
stratum, usually by random, or systematic sampling methods.

Example 1: In a certain study, 200 samples are taken from the population of 50,000
individuals. The population is divided into strata based on their schools. Using
stratified sampling, we have:
Distribution of Percentage from the Sample Units
Strata
Population Population Per Stratum
UST 15,000 30% 60
UP 10,000 20% 40
NU 25,000 50% 100
Total 50,000 100% 200

Example 2: Using stratified random sampling, select a sample of 387 students from the
population which are grouped according to the cities they come from. The table
below shows the number of students per city.

City Population (N)


A 13,000
B 10,500
C 8,000
D 5,000

Solution:
To determine the number of students to be taken as sample from each city, we divide the
number of students per city by the total population (which is 36,500) and multiply the result
by the total sample size (which is 387).

City Population (N) Sample (n)


13,000
A 13,000 (387) = 137.84 or 138
36,500
10,500
B 10,500 (387) = 111.32 or 111
36,500
8,000
C 8,000 (387) = 84.82 or 85
36,500
10,500
D 5,000 (387) = 53.01 or 53
36,500

Situational Example:
Some population can be divided into groups called strata. Ideally, strata are
homogeneous groups of population units, that is, units in a given stratum are similar in some
important characteristic, whereas those in different strata differ in those characteristics. For
example, students in a university could be grouped by their year or major, or a city could be
divided into geographics zone. If a population is divided into homogeneous strata and a simple
random sample is selected from each stratum, the result is a stratified random (see Figure 1.b).
Suppose a company receives its supply of switches from three different suppliers (say, A, B,
and C).

Then each supplier forms a natural stratum. By taking separate random samples from
different suppliers (strata), the industrial engineer can also measure different among suppliers.
Such a stratified sampling procedure could provide an accurate estimate within each stratum
while leading to a more precise (reduced variation) combine estimate for the population as a
whole.
Figure b.1

Researchers obtain stratified samples by dividing the population into groups (called
strata) according to some characteristic that is important to the study, then sampling from each
group. Samples within the strata should be randomly selected. For example, suppose the
president of a two-year college ants to learn how students feel about a certain issue.
Furthermore, the president wishes to see if the opinions of the first-year students differ from
those of the second-year students. The president will randomly select students from each group
to use in the sample.

Strata (plural of stratum) – groups or classes inside a population that share a common
characteristic
Example:
In the population of all undergraduate college students, some strata might be freshmen,
sophomores, juniors, or seniors. Other strata might be men or women, in-state students or
out-of-state students, and so on.

In the method of stratified sampling, the population is divided into at least two distinct strata.
Then a (simple) random sample of the certain size is drawn from each stratum, and the
information obtained is carefully adjusted or weighted in all resulting calculations.

The groups or strata are often sampled in proportion to their actual percentages of occurrence
in the overall population. However, other (more sophisticated) ways to determine the optimal
sample size in each stratum may give the best results.

3. Cluster sampling
This technique is frequently applied on geographical basis when the population from which a
sample is to be selected includes heterogeneous groups.

Cluster sampling sometimes called area sampling, it is applied on a geographical basis. It is


geographical basis. It is generally done by first sampling at the higher levels before going down to the
lower levels. For instance, samples are randomly taken from provinces first; then subsequent cities,
the municipalities or barangays, and then households.
Situational Example:
If the population items are packed 24 to a box, it would be time-consuming to take a random
sample or a systematic sample that basically involves opening each box. We can consider each box as
one cluster, select a few boxes at random from the entire shipment, and then inspect all items from
the selected boxes. This sampling method is known as cluster sampling (see Figure 3.1).

ln cluster sampling, a population is divided into groups called clusters, and a simple random of
cluster Is selected. Then, either every unit is measured in each sampled cluster, or a second-stage
random sample is selected from each sampled cluster. Cluster sampling methods may seem similar to
stratified random sampling. In stratified sampling, a population is divided into homogeneous strata,
whereas in cluster sampling, clusters are often nonhomogeneous. In stratified sampling, all strata are
sampled, whereas in cluster sampling a census is conducted or random samples are taken from
randomly selected clusters. Cluster sampling is often a cost-saving method because a sampling frame
is not needed for each cluster, but only for the selected clusters. It is particularly cost saving for
geographically dispersed clusters, such as plots of trees in the forest or rural towns scattered across a
country.

Figure 3.1

Researchers also use cluster samples. Here the population is divided into groups called clusters
by some means such as geographic are or schools in a large school district, etc. Then the researcher
randomly selects some of these clusters and uses all members of the selected clusters as the subjects of
the samples. Suppose a researcher wishes to survey apartment dwellers in a large city. If there are 10
apartment buildings in the city, the researcher can select at random 2 buildings from the 10 and
interview all the residents of these buildings. Cluster sampling is used when the population is large or
when it involves subjects residing in a large geographic area. For example, if one wanted to do a study
involving the patients in the hospitals in New York City, it would be very costly and time-consuming
to try to obtain a random sample of patients since they would be spread over a large area. Instead, a
few hospitals could be selected at random, and the patients in these hospitals would be interviewed
in a cluster.

Cluster sampling is a method used extensively by government agencies and certain private
research organizations. In cluster sampling, we begin by dividing the demographic area into sections.
Then we randomly select sections or clusters. Every member of the cluster is included in the sample.
For example, in conducting a survey of school children in a large city, we could first randomly select
five schools and then include all the children from each selected school.

The four basic sampling methods/techniques are summarized in Table A below.


Table A Summary of Sampling Methods
Subjects are selected by random numbers.
Random
Use a simple random sample from the entire population.
Subjects are selected by using every kth number after the first subject is randomly
selected from l through k.
Systematic
Number all members of the population sequentially. Then, from a starting point
selected at random, include every kth member of the population in the sample.
Subjects are selected by dividing up the population into groups (strata), and subjects
are randomly selected within groups.
Divide the entire population into distinct subgroups called strata. The strata are
Stratified
based on a specific characteristic such as age, income, education level, and so on.
All members of a stratum share the specific characteristic. Draw random samples
from each stratum.
Subjects are selected by using an intact group that is representative of the
population.
Cluster Divide the entire population into pre-existing segments or clusters. The clusters are
often geographic. Make a random selection of clusters. Include every member of
each selected cluster in the sample.

4. Non-random Sampling
In this technique, not all the population has equal chance to be selected. The selection is
influenced by the goal of the researcher. There are three forms of non-random sampling:

a. Purposive Sampling. The samples are chosen based on purpose or certain criteria. For
example, in a population of college students, you are studying the effects of being sporty in
the academic performance. The possibility is that you will choose the athletes of the
university by purpose. Or let us say for example, you want to know the modern ways of
college students in courting a girl. You will need to choose the male part of the population.
A purposive sample, also referred to as a judgmental or expert sample, is a type of
nonprobability sample.
Purposive sampling is used in very small sizes. For example, this can be used if the subjects
of the study are deans of certain universities or area managers of certain institutions.

Situational Example:
Some procedures used probabilistic approaches to determine which item of population is to
be selected in the sample, and some don’t. For example, although the selection of a jury
pool may be random, the actual jury selection from the available pool of jurors is not a
random process. Using their judgement, lawyers from both parties question prospective
jurors and decide who is to be selected as a jury for a given case. This may result in biased
jury, that is, selection of those with specific opinions. Such a procedure is known as
judgmental sampling.
This approach uses no probability and is entirely based on the judgment of the person
selecting the sample. Judgmental sampling is not necessarily always a bad technique, at times,
it can be advantageous. For example, a team of environmental engineers is checking a site
for possible chemical pollution int the soil. Instead of randomly determining the spots from
which the soil samples are taken, the engineers use their expertise in how pollution spreads
and their knowledge about the layout of that particular site to select the most likely spots
for testing. This could even achieve the required results with a smaller sample size, thus
reducing the case of the study.

b. Quota-Sampling. In this technique, a certain limit is pre-established to determine who among


the population can be part of the sample. A good example of this is the determination of
the students who can qualify in a university. An admission test will be the basis and a quota
rule is already stated even before the exam. For example, the top 5% of the examinees shall
be admitted by the university.

Quota sampling is defined as a non-probability sampling method in which researchers create


a sample involving individuals that represent a population. Researchers choose these
individuals according to specific traits or qualities. They decide and create quotas so that the
market research samples can be useful in collecting data. These samples can be generalized
to the entire population. The final subset will be decided only according to the interviewer’s
or researcher’s knowledge of the population.

For example, a cigarette company wants to find out what age group prefers what brand of
cigarettes in a particular city. He/she applies quotas on the age groups of 21-30, 31-40, 41-
50, and 51+. From this information, the researcher gauges the smoking trend among the
population of the city.

Quota sampling includes a specified number of persons of certain types to be taken as sample.

Types of quota sampling:


Quota sampling can be of two kinds – controlled quota sampling and uncontrolled quota
sampling. Here’s what they mean:

• Controlled quota sampling:


Controlled quota sampling imposes restrictions on the researcher’s choice of samples.
Here, the researcher is limited to the selection of samples.

• Uncontrolled quota sampling:


Uncontrolled quota sampling does not impose any restrictions on the researcher’s choice
of samples. Here, the researcher chooses sample members at will.

Quota sampling example:


A researcher wants to survey individuals about what smartphone brand they prefer to use.
He/she considers a sample size of 500 respondents. Also, he/she is only interested in
surveying ten states in the US. Here’s how the researcher can divide the population by
quotas:

Gender: 250 males and 250 females


Age: 100 respondents each between the ages of 16-20, 21-30, 31-40, 41-50, and 51+
Employment status: 350 employed and 150 unemployed people.
(Researchers apply further nested quotas. For eg, out of the 150 unemployed people, 100
must be students.)
Location: 50 responses per state

Depending on the type of research, the researcher can apply quotas based on the sampling
frame. It is not necessary for the researcher to divide the quotas equally. He/she divides the
quotas as per his/her need (as shown in the example where the researcher interviews 350
employed and only 150 unemployed individuals). Random sampling can be conducted to
reach out to the respondents.

How to perform quota sampling:


Probability sampling techniques involve a significant amount of rules that the researcher
needs to follow to form samples. But, since quota sampling is a non-probability sampling
technique, there are no rules for formally creating samples. Usually, there are four steps to
form a quota sample. Here are the steps:

• Divide the sample population into subgroups: With stratified sampling, the researcher
bifurcates the entire population into mutually exhaustive subgroups, i.e., the elements of
each of the subgroups becomes a part of only one of those subgroups. Here, the researcher
applies random selection.

• Figure out the weightage of subgroups: The researcher evaluates the proportion in which
the subgroups exist in the population. He/she maintains this proportion in the sample
selected using this type of sampling method.
For example, if 58% of the people who are interested in purchasing your Bluetooth
headphones are between the age group of 25-35 years, your subgroups also should have
the same percentages of people belonging to the respective age group.

• Select an appropriate sample size: In the third step, the researcher should select the sample
size while maintaining the proportion evaluated in the previous step.
If the population size is 500, the researcher can pick a sample of 50 elements.
The sample chosen after following the first three steps should represent the target
population.

• Conduct surveys according to the quotas defined: Make sure to stick to the predefined
quotas to achieve actual actionable results. Don’t survey quotas that are full and focus on
completing surveys for each quota.

c. Convenience Sampling. The sample is selected based on the accessibility of the researcher in
convenience sampling. For example, the researcher is doing a study about the performance
of universities in the Philippines. If the researcher lives near Manila, he has the option to take
the universities in Manila as samples.

In this sampling, a researcher uses subjects that are convenient. For example, the researcher
may interview subjects entering a local mall to determine the nature of their visit or perhaps
what stores they will be patronizing. This sample is probably not representative of the
general customers for several reasons. For one thing, it was probably taken at a specific time
of day, so not all customers entering the mall have an equal chance of being selected since
they were not there when the survey was being conducted. But convenience samples can be
representative of the population. If the researcher investigates the characteristics of the
population and determine that the sample is representative, then it can be used.

Convenient sampling is most convenient and fastest sampling technique that make use of the
telephone, mobile phones, or the internet.

Situational Example:
Instead of using his expertise to select spots for getting soil sample, suppose an engineer takes
soil samples from the easily accessible areas. Then the sample may not give representative
results for the entire location. In fact, the result might be misleading due to missing some
highly polluted spots in not-so-easy-to-reach areas. This sampling methods results in a sample
of convenience, another method that results in biased outcomes.

Convenience sampling simply uses results or data that are conveniently and readily obtained.
In some cases, this may be all that is available, and in many cases, it is better than no
information at all. However, convenience sampling does run the risk of being severely
biased. For instance, consider a newsperson who wishes to get the “opinions of the people”
about a proposed seat tax to be imposed on tickets to all sporting events. The revenues from
the seat tax will then be used to support the local symphony. The newsperson stands in front
of a concert hall and surveys the first five people exiting after a symphony performance who
will cooperate. This method of choosing a sample will produce some opinions, and perhaps
some human interest stories, but it certainly has bias. It is hoped that the city council will not
use these opinions as the sole basis for a decision about the proposed tax. It is good advice
to be very cautious indeed when the data come from the method of convenience sampling.

d. Volunteer sampling
Volunteer sampling also results in biased outcomes. This method is popular with radio and
television shows for conducting quick opinion polls. In volunteer sampling, it is up to
individuals to decide whether they want to be part of the survey.

Example:
A company was viewing Mobile, Alabama, as a possible site for building a liquefied natural
gas (LNG) processing center. Some residents were opposed to the idea, arguing that it poses
a high risk to lives. A local television station conducted a survey to address the issue of
whether Mobile should give permission to build a LNG processing center in the Mobile area.
A telephone number was provided and respondents were asked to call and register their
opinions. A computer responded to the calls and prompted callers to press 1 if they supported
building the LNG center in a Mobile area and 2 if they were against it. The television station
then counted number of the 1’s and 2’s to determine the support for and the opposition to
the LNG center. But they might have inadvertently introduced bias into the results, because
only those who feel strongly about the issue (either in support or in opposition) are likely
to call the advertised number and register their opinion.
e. Accidental sampling
Only those whom the researcher meets by chance are included in the sample when using this
technique.

A sampling method is biased if it tends to give samples in which some characteristic of the population
is underrepresented or overrepresented.

Biased sampling methods result in values that are systematically different from the population values,
or systematically favor certain outcomes, resulting in systematic overestimation or underestimation.
Purposive Sampling or judgmental sampling, quota-sampling, convenience sampling or samples of
convenience, volunteer sampling, and accidental sampling are some of the methods that generally
result in biased outcomes. Sampling methods, such as simple random or probability sampling, stratified
random sampling, systematic sampling, and cluster sampling, that are based on probabilistic selection
of samples generally results in unbiased outcomes.

All sampling designs are subjects to sampling error, variation inherent in the sampling process. Even if
the survey is repeated using the same sample size and the same questionnaires, the outcome will be
different. This sampling error can be controlled to some extend by a good sampling design and can
be measured by the inferential techniques.

There are other types of errors, however, that are not so easily controlled or measured. One major
problem in a sample survey is use of an incorrect sampling frame.

A telephone survey conducted using a telephone directory exclude those with unlisted phone
numbers; however, computerized random digit dialing solves this problem. Another problem that
arises in sample surveys is that interviewers sometimes substitute another person if selected person is
not available, an action that can introduce serious bias into the results. Under time and money
constraints, reliable and accurate results are more likely to be forthcoming from a small sample survey
with a good design than from a large survey that is hastily put together.

Sources of Bias in Surveys


When any controversial event occurs, newspapers, television stations, and radio stations, rush to
conduct opinion polls and report results of these surveys. However, as The Wall Street Journal reporter
Cynthia Crossen (1992) notes, “These are most likely to be wrong because questions are hastily drawn
and poorly pretested, and it is almost impossible to get a random sample in one night.” It is possible
that those people inaccessible on one night have systematically different opinions about the issue than
those accessible people.

For a survey to produce reliable results,


1. The survey must be properly designed and conducted.
2. Samples must be selected using proper randomization techniques. Nonrandom selection of the
sample will limit the generalizability of results.
3. The interviewers must be trained in proper interviewing techniques to avoid biases in responses.
The attitude and behavior if the interviewer should not lead to any specific answers resulting in
biased outcomes.
4. Wording of the question must be selected carefully so that leading questions and other forms of
biased question can be avoided, as wording of the questions affects the response.
A survey is biased if it systematically favors certain outcomes. A bias can occur for so many different
reasons, such as improper wording of a question, the interviewer’s behavior, nonrandom selection of
the sample, improper frame, and so on.

To estimate the percent of high school children that smoke, children are surveyed from one high
school and asked whether they smoke cigarettes. If a parent is present at such an interview, the results
will be biased because students may systematically tend to favor a negative answer over a positive
one regardless of the truth. Such a bias that occurs due to the behavior of the interviewer or respondent
is known as the response bias. It is possible to reduce response bias by carefully training interviewers
and supervising the interview process. A nonresponse bias may occur if the person selected for the
interview cannot be contacted or refuses to answer.

Most products come with a warranty card that a purchaser is expected to fill out and send to the
manufacturer. One manufacturer decides to conduct a customer satisfaction survey, and he selects
customers from the list of warranty cards received. By doing so, those choose not to send the warranty
card back or forgot to do so were totally excluded from the survey. If there is systematically difference
in the opinions of those who chose to send the warranty card back and those who chose not to send
it back, then the results will be biased. This type of bias is known as an undercoverage bias, and it may
occur if part of the population is left out of the selection process. For example, in a telephone survey,
individuals without a telephone are left out of the selection process. In the United States, almost 98%
of households have telephones, and a very small percent of population is left out of the selection
process. But in many African countries, for example, less than 4% of households have telephones, and
a telephone survey would give seriously biased results. In every census, a certain percent of population
is missed due to undercoverage. The undercoverage is higher in poorer sections of larger cities due to
nonavailability of addresses.

A wording effect bias may occur if confusing or leading questions are asked. For example, an
interviewer may first state, “The American Dental Association recommends brushing your teeth three
times a day,” and then ask, “How often do you brush your teeth every day?” To avoid looking bad,
the respondent may feel compelled to answer “three or more times” even if he or she does not brush
that often. The responses are likely to be higher number than that the truth represents. In this situation,
the wordings effect bias could be reduced or avoided by simply asking, “How often do you brush
your teeth every day?”

Critical Thinking

Sampling frame We call the list of individuals from which a sample is actually
Selected the sampling frame. Ideally, the sampling frame is the
entire population. However, from a practical perspective, not all
members of a population may be accessible. For instance, using a
telephone directory as the sample frame for residential telephone
contacts would not include unlisted numbers.

Undercoverage When the sample frame does not match the population, we have
what is called undercoverage. In demographic studies,
undercoverage could result if the homeless, fugitives from the law,
and so forth are not included in the study.

A sampling frame is a list of individuals from which a sample is


actually selected.

Undercoverage results from omitting population members from


the sample frame.

Sampling error In general, even when the sampling frame and the population
match, a sample is not a perfect representation of a population.
Therefore, information drawn from a sample may not exactly
match corresponding information from the population. To the
extent that sample information does not match the corresponding
population information, we have an error, called a sampling error.

A sampling error is the difference between measurements from a


sample and corresponding measurements from the respective
population. It is caused by the fact that the sample does not
perfectly represent the population.

Nonsampling error A nonsampling error is the result of poor design, sloppy data
collection, faulty measuring instruments, bias in questionnaires, and
so on.

Sampling errors do not represent mistakes! They are simply the


consequences of using samples instead of populations. However,
be alert to nonsampling errors, which may sometimes occur
inadvertently.

Sample Size Formula


Most of the time population is used simultaneously with sample but these two are actually
different terms. Population is the complete set of individuals or subject while sample is just a
representative of the whole population. The elements in a sample are drawn from the population.

How to get the sample size?


𝑁
n=
1+𝑁𝑒 2

Where: n = sample size


N = population
e = margin of error
This formula is known as Sloven’s formula.
Example 1: A researcher wants to conduct a study in a university with 10,000 students. If he wants to
achieve 90% precision, how many students must he take as his sample?

Given: N = 10,000; e = 10%

𝑁 10,000
Solution: n = = = 99 students
1+ 𝑁𝑒 2 1+(10,000)(0.102 )

Example 2: A researcher wants to study the academic performance in Mathematics of students in a


certain school. The school has a population of 12,000 students. If the researcher allows a
margin of error of 5%, how many students must he include in his sample?

Solution: It was given that N = 12,000 and e = 5% or 0.05. Using Sloven’s formula, the
sample size is computed as follows:

𝑁
n=
1+𝑁𝑒 2
12,000
=
1+12,000(0.05)2
12,000
=
1+12,000(0.0025)
12,000 12,000
= =
1+30 31
= 387.096 or 387

Thus, the researcher must take 387 students as his sample.


Exercise 2.1
Sampling Data

Solve for what is asked in each question. Show your given and solution.
1. How many sample units must be obtained from 7,000 employees, if 10% margin of error is used?
2. Find the margin of error of the research conducted by DOST about the people affected by the
Mayon Volcano eruption if a sample off 1875 people is used over a population of 22,750.

Exercise 2.2
Sampling Technique

Solve for what is asked in each question. Show your given and solution.
1. In a population of 30,000 Chinoy, the researcher wants to select 315 samples to be interviewed. If
the researcher aims an 85% precision, and wants to use the systematic sampling, the every kth
individual must be taken from the population. Find k.
2. Determine the number of samples that can be selected from a population of 17,500 people if every
301st person is taken as part of the sample.

Exercise 2.3
Check your Progress

1. Find the sample size required from a population of 10,000, given a margin of error of 5%. Use
Sloven’s formula.
2. Distribute the sample size obtained in item 1 as classified to the following strata:

Class A 500
Class B 2,500
Class C 3,000
Class D 4,000

Exercise 2.4
(Applying the Concepts)

American Culture and Drug Abuse


Assume you are a member of the Family Research Council and have become increasingly concerned
about the drug use by professional sports players. You set up a plan and conduct a survey on how people
believe the American culture (television, movies, magazines, and popular music) influences illegal drug
use. Your survey consists of 2250 adults and adolescents from around the country. A consumer group
petitions you for more information about your survey. Answer the following questions about your
survey.
1. What type of survey did you use (phone, mail, or interview)?
2. What are the advantages and disadvantages of the surveying methods you did not use?
3. What type of scores did you use? Why?
4. Did you use a random method for deciding who would be in your sample?
5. Which of the methods (stratified, systematic, cluster, or convenience) did you use?
6. Why was that method more appropriate for this type of data collection?
7. If a convenience sample were obtained consisting of only adolescents, how would the results of
the study be affected?

You might also like