Research
Methodology and
Statistical Tools
What is the right research
design for the study?
Most researches use descriptive research design because it is simple,
requires minimal explanation, and does not involve complex
statistical testing like hypothesis testing. It is useful in both
academic and applied business studies, though it can still be
overwhelming if many activities need to be described. However,
when measurement and analysis go beyond what descriptive
research can cover, other designs—such as correlational and
causal research—are used, which are more complex and require
higher research skills.
Guidelines for using the Correlational
Research Tools
1. If the problem involves hypothesis testing, then decide on:
a. Hypothesis of relationships (between variables) – use Pearson R Correlation (for parametric
data) or Kendall’s Tau B, Spearman’s Rho, Chi square (for non-parametric data).
b. Hypothesis of difference - use:
b.1 Paired T-Test – for comparing a pair of variables (interval scaled or ordinal) 18
b.2 Independent Samples T-Test – for comparing a variable (interval or ratio scaled) with a nominal
variable with 2 categories only.
b.3 One-Way ANOVA – for comparing a variable (interval or ratio scaled) with a nominal variable
with at least 3 or more categories.
Guidelines for using the Causal
Research Tools
1. If the problem is to test the existing theoretical framework and validate its measurement
variables for research application, then the right techniques is Confirmatory Factor Analysis (CFA).
CFA is also a very good validation tool.
2. If the problem is complex and characterized by complex relationships (which cannot be solved
by either of the following: multiple regression analysis, factor analysis, cluster analysis, and
conjoint analysis), then the right technique is either: Covariance based-Structural Equation
Modeling (CB-SEM) or Partial Least Squares-Structural Equation Modeling (PLS-SEM).
proper sampling method
When conducting research, students and researchers must carefully choose the right
sample size to make their results reliable. Some schools even set standard sample sizes for
students to follow. The required sample size depends on two things: the margin of error
(how much error you are willing to allow) and the confidence level (how sure you want to be
about the results).
For instance, undergraduate research commonly uses around 100 respondents, which
corresponds to a 9.78% margin of error at a 68.3% confidence level. This is less precise, but
acceptable at the undergraduate level where studies are smaller in scope. In contrast,
graduate research usually requires around 377 respondents, which gives a 5% margin of
error at a 95% confidence level. This ensures more accuracy and reliability
appropriate sample size
As to the appropriate sample size, it depends on the school, the panel of
examiners, or the contracted researcher as to the appropriate sample
size. It should however be 100 or more. The problem with a big sample
size is the cost and time involved: it becomes so expensive and it
requires a much longer period of time to gather the data.
Other sampling size formula that the
students or researchers can apply
1. Slovin’s Formula (if the population is known)
. 2. Cochran’s Formula (if the population is unknown)
Random Sampling or Purposive
Sampling?
If the objective of the study is to generalize the result across the
population, then use random sampling. If the objective of the study is to
determine the target market for a particular organization and is not
necessary to generalize it across the population (since the purpose is like
a case study), then use purposive sampling.
Purposive sampling
Purposive sampling is a non-probability sampling method where the researcher deliberately
chooses respondents who meet certain criteria relevant to the study. The process begins by
qualifying the respondents—that is, checking if they fit the requirements (for example, only
business owners, teachers, or patients with a specific condition). After identifying those who
qualify, the researcher then selects participants using a quota method, meaning they take only a
specific number of respondents from each group until the desired sample size is reached.
Convenience sampling is generally not appropriate for research because it involves collecting data
from whoever happens to be available at the time of the survey, without considering their
qualifications or following a proper sampling frame.
random sampling
Random sampling is a method of selecting participants for a study where
every individual in the population has an equal chance of being chosen. It is
one of the most reliable sampling techniques because it reduces bias and
makes the sample more representative of the whole population.
Types of Random Sampling
1. Simple Random Sampling – if the population is known
2. Stratified Random Sampling – if the population is known
3. Random Sampling based on Poisson distribution – if the population is
not known