UNDERSTANDING DATA

AND SYSTEMATIC WAYS

TO COLLECT DATA
Research
Design
 The Research Design

 The research design is a plan that presents how the researcher

intends to study an empirical question (Johnson and Reynolds
2012).
 Two of the most common quantitative research designs used in
social science research are the descriptive research design and
correlational research design.

The Research Design

 Descriptive Research Design

Descriptive research “paints a picture “ with words

or numbers (Neuman 2007, 364).

Studies that adopt as descriptive research design

puts emphasis on collecting, organizing, and
presenting the target population through primary
data gathering using surveys, field observations,
interview, etc.
It may also use secondary data gathering
approaches such as the use of databases or
secondary sources such as datasets produced in
other studies.

The main method of analysis used in descriptive

research design is descriptive statistics.

This means is typically uses measures such as

the mean, the mode, and the median to describe
data.

Descriptive Research Design

 Correlational Research Design

 Correlational research is used to establish or explore

relationships, associations, or interdependence between two
or more aspects of a situation (Kumar 1996).
 It is generally concerned with empirically establishing and
statistically measuring the nature, degree, and direction of
relationships between two or more variables of interest using
techniques of statistical analysis.
 A correlational research design is appropriate for studies
examining associations to explain how changes in one
variable correlate with the values of another.
 The primary method of analysis in correlational
research design is hypothesis-testing which is
center to inferential statistics.

 Analysis uses bivariate analysis such as chi-square

test of independence , chi-square goodness of fit
test, person’s correlation, and simple liner
regression.

 It may also use or more complex forms of

multivariate analysis such as multiple regression
and panel regression analysis.
Correlational Research Design
Sampling
There are four types of
probability sampling frames:
▪ Simple Random Sampling
▪ Systematic Random Sampling
▪ Stratified Random Sampling
▪ Clustered Sampling
Simple Random Sampling

▪ In simple random sampling (SRS),

each element of the population has
an equal chance of being selected.
Systematic Random Sampling
The systematic sampling technique, a list of
all the elements in the population is also
prepared but instead of using a random
numbers table to draw the sample, a
sampling interval is used.

The formula in computing the sampling

interval is K=N/n (population size/sample
size). Meaning, you select every kth element
in your population list.
Stratified Random Sampling
▪ Stratified random sampling reduces the
sampling is organizing the population into
homogenous subsets or grouping the
elements in the population according to
shared characteristic/s.
▪ The researcher selects the appropriate
number of elements for each identified
subset or group, ensuring that each selected
element is proportionally represented in the
total population.
Stratified Random Sampling

▪ This is done by computing for the

sampling fraction –the desired size of
the sample divided by the population
size, or n/N.
Clustered Sampling

▪ Clustered sampling is a multistage

method adopted when it is either
impossible or impractical to compile an
exhaustive list of the elements found in
the target population.
Clustered Sampling

▪ Naturally, population consists of “natural

groups,” or clusters which are initially
sampled . Then the members of each
group are subsampled afterward. This
involves the repetition of two sampling
steps: listing and sampling.
There are four types of non-
probability sampling frames:
▪ Purposive Sampling
▪ Convenience Sampling
▪ Quota Sampling
▪ Snowballing Sampling

▪ Compared to probability sampling, non-

probability sampling follows a sample
selection based on your subjective
judgment.
Purposive Sampling

▪ Purposive sampling is guided by your

main purpose or reason for conducting
the study. Thus, you select the
elements you think are useful for your
study.
Purposive Sampling

The main goal of purposive sampling is

not to generalize, but to focus on
particular characteristics or subjects
which you believe are capable of
providing answers to the research
questions.
This sampling technique is also called
as “judgmental sampling” or “subjective
sampling.”
Convenience Sampling

▪ Elements selected through

convenience sampling are those
easily accessible to you.
Quota Sampling
▪ Quota sampling is often likened to
proportionate or stratification of the
population. Sampled elements should
be proportionate to their
representation in the population.
Follow these three basic steps in creating a quota
sample.
First, select the predetermined characteristics (or
stratification) and divide the population accordingly.

Second, calculate the quota needed for each

stratum.

Third, invite subjects until the quota per stratum is

attained.
.

Quota Sampling
Snowball Sampling
▪ An appropriate nonprobability
sampling technique to adopt is the
snowball sampling where the data is
collected from few known members of
the target population.
▪ These known members will then be
asked to provide information that will
help locate other members of the
desired population.
Designing
Survey
Instrument
 Survey and Questionnaires
 Surveys often take the form of
questionnaires –a structured set of
questions administered to the respondents
personally, by telephone, by mail, or by
electronic means.
 It is pursued when the population is too
large to be observed or interviewed.
 Survey and Questionnaires
 Surveys are only useful when the unit of
analysis used are individuals or persons.

 Surveys allow the researcher to obtain

sensitive information once the respondents are
assured of confidentiality and anonymity.
Basic Principles in Designing Survey Questionnaires
 Constructing and administering surveys is
subjected to standards and ethics, balancing
between your need to get the maximum amount
of information and from disenfranchising the
respondent.
 One factor respondents weigh when answering a
survey is the questionnaire length. Some
respondents might experience what is called
“survey fatigue” (Johnson and Reynolds 2011,
320) for questionnaires that are too long.
Basic Principles in Designing Survey Questionnaires

 Surveys must also be clear and unambiguous.

Respondents must be able to quickly understand
the purpose of the survey and provide answers
without difficulty.

 Double-barreled questions, or those that

have multiple parts, should also be avoided.
Basic Principles in Designing Survey Questionnaires

 Critically examine your questions when the

word “and” appears in a particular item.

 Similarity, unfamiliar terms and jargon may

lead to confusion and misinterpretation.
Whether possible, always substitute
clearly defined, familiar terms.
Basic Principles in Designing Survey Questionnaires

 Pretesting helps you detect problems in the

questionnaire and the survey administration
process.
 This may be done through focused group
discussions in which the questionnaire is
reviewed by a selected panel.
 This allows the identification of ambiguous
items in the questionnaire. This helps establish
the variables or indicators of interest
Basic Principles in Designing Survey Questionnaires

 Pilot testing is a pretesting method in which

you test your survey questionnaire in the
same circumstances you intend to carry it out.
 After pilot-testing, it is recommended that you
share your findings and experiences with
others to enhance knowledge and research
process (Colton and Covert 2007).
Types of Survey Questionnaire
 Open-Ended Question
The respondents are allowed to answer the
questions freely.
Responses must be recorded verbatim –
especially because coding and analysis will
rely on the subject’s exact responses.
Open-ended questions often need probing
or follow-typically ask to “how” and “why”
of something.
 Dichotomous Question
Dichotomous question have two possible
answers, often either yes/no, true/false, or
agree/disagree.
These questions are used when the
researcher wants to clearly distinguish the
respondent’s opinion, preference, experience
or behavior.
 Multiple-Response Question

There are certain questions that necessitates

the respondents to provide more than one
answers.
 Matrix Question

There are instances where a number of

questions you intend to ask have the same set
of possible answers. Thus, it is possible to
construct a matrix of items and answers for
the sake streamlining the survey.
 Contingency Question

Contingency questions are intended for

certain respondents only, depending on the
provided answers.
Planning for Data
Analysis
Different Ways of Analyzing Quantitative Data
1.Descriptive Statistics
Descriptive statistics summarizes and
describes the characteristics of a sample or
the relationship among variables in a sample.
It helps present quantitative data in
manageable, easily understandable forms
(Babbie 2010).
Descriptive statistics uses various summary
measures and graphing techniques that help
organize and summarize data for easy
understanding.
There are two ways to present data in descriptive
analysis: the measure of central tendency and
measure of spread.

• Measure of spread are the ways of summarizing

a group of data. This central tendency can be
described using the mean, median, and mode.
• Measure of spread are the ways of summarizing
a group of data by describing how spread out
the data is.
2. Inferential Statistics

Inferential Statistics is the branch of statistics that

makes relevant inferences about the larger
population from a smaller population.
This statistical method is based on inductive
reasoning, which tests whether what has been
observed true for the smaller sample in study is
also true for the larger population (Sirkin 1999).
Inferential statistics moves beyond describing
and makes assertions beyond the immediate
data.
Writing the Research
Methodology
 Writing the Research Methodology
The methodology of a research paper
describes all the detailed steps you intend
to do in order to systematically collect and
analyze the data you need to answer your
research question.
 Its primary purpose is to guide
researchers in data collection and
analysis.
Defining and Operationalizing the Key
Concepts and Variables
 In the context of quantitative research,
operationalizing a concept involves
adopting a specific working definition that
guides how data is observed, measured,
and collected.
 In most cases, the process of identifying
indicators is one step in the
operationalization of a concept. Indicators
are a set of criteria reflective of the
concept.
Determining the Level and Unit of Measurement
 The type, level and unit of measurement should be established in
the process of operationally defining concepts.
 It is important that all key concept and variables in a research
project have been operationally defined before planning for the
data collection methods.
 These definitions determine the specific pieces of information
that needs to be collected.
 They tell us exactly what specific information we need to look
for and gives us an idea of how to find them.
 It serves as a guide for designing instruments used for data
gathering, such as survey questionnaires and observation guides.
 Determining the Level and Unit of
Measurement

 Quantitative research involves classifying

variables as either categorical or continuous, or
as nominal, ordinal, interval, or ratio.
 In addition to the level of measurement, the
researcher should establish at what level you
collect data (information at the individual , family,
barangay, organization, city, country level, etc.)
and how many observations they need.
 These are referred as the units of observation or
analysis.
 Identifying the population and Choosing
a Sampling Frame
◂ An essential part of data collection is
identifying the population from which
you will draw a sample the
population of the Philippines), people
(e.g.,Filipinos, college students),
cities (e.g., in Metro Cebu),
organizations (e.g., Catholic
Educational Association of the
Philippines or CEAP), and groupings
of individuals (e.g., all Grade 11
students of a school.
 Choosing the Method for Collecting Data
 Primary data pertains to first hand collected by
you.
 Primary data collection methods include surveys,
observations, experiments, and content analysis,
among others.
 Secondary data pertains to second hand
information that the researcher collects from
sources produced by others.
 Secondary data collection methods on the other
hand include collection of information from
databases all literature research among others.
 Outlining the Plan of Analysis
 The plan of analysis outlines how
you intend to analyze the data you
have collected.
 It is developed around the research
questions and the framework of your
research project.
 In thinking your plan of analysis,
carefully consider how the data can
be analyzed in a way that will allow
you to answer the research
questions.