SKPD 8124 Research

Methodology
Lecture 3: Basic Statistics
for Research
Ts. Dr. Mohammad Lui Juhari
Basic Statistics
for Research
Content
Introduction

Descriptive Statistics

Inferential Statistics
Introduction
vData analysis can be said as method of applying
the correct statistical or logical
techniques/methods to the raw data collected in
a research and turn it into something which
makes sense.

vData analysis is the most important part in any

research.

vIt is done to summarize collected data and

findings of a research .
Introduction

vTo do the analysis, researcher need to

complete data collection to obtain raw data.

vThen, use statistical analysis (descriptive or

inferential) to identify patterns, relationships
or trend in collected data.
Introduction

vThere are variety of statistical methods and

testing can be apply into a research but
analysis must follow objectives of the research
as well as answer and prove the hypotheses.

vOutcome of these methods usually will be

reported in the form of graphs, figures and
table as research won’t be reporting raw data.
Introduction

vThere are various software and application

can be used to perform analysis instead of
using manual calculation.

vThese include Microsoft Excel, Tableau,

MATLAB, SPSS, AMOS, SmartPLS etc.
 Introduction

Statistics

Descriptive
Inferential statistics
statistics
 Statistical Analysis
There are three types of statistics.

What is the difference between these three????

Examples

§ Univariate – focus on one variable at a time

§ Bivariate – focus on two variables at a time and can be presented as
contingency table
§ Multivariate – have multiple variables at a time

Univariate Bivariate Multivariate

 Descriptive Statistics
Contingency table
 Descriptive Statistics
v Descriptive statistics or analysis done to summarize and organize
characteristics of a set of data collected in the research.

v Before any statistical analysis being performed, it is required to describe

characteristics of responses, demographic variables and other raw data.
 Descriptive Statistics
v There are three main types of descriptives statistics

1. The distribution - frequency of each value

2. The central tendency – averages of value
3. The variation or dispersion - how values are spread out
Descriptive Statistics
Descriptive
statistics

Measure of
Measure of
Distribution central
tendency variation

Mean Range

Standard
Median
deviation

Mode Variance

Interquartile
range
 Descriptive Statistics
v Frequency distribution:

§ Can be summarize in table or graph, the form of numbers (frequency) or

percentage

§ For example, you have 400 respondents in a survey on product A. You want to
see the frequency of responses based on gender or age groups
 Descriptive Statistics

Variables Frequency (n) Percentage (%)

Gender Male 125 31.25
Female 275 68.75
Age group 21 – 30 years old 280 70
31 – 40 years old 89 22.25
41 – 50 years old 31 7.75
 Descriptive Statistics
v Measures of central tendency:

§ Involve estimating the center/average of a data set usually using mean, median
and mode
Descriptive Statistics
 Descriptive Statistics
v Measures of variation:

§ Give indication on how a set of data is being spread out through the findings.

§ Mostly presented using range, standard deviation and variance.

§ Range for examples tell you of how far apart the most extreme response score
or minimum and maximum value are spread out.

Minimum value Maximum value

Let say 10 students score 60, 64, 67, 74, 76, 78, 80, 83, 90, 95 in mathematic

Range of score 95 – 60 = 35
 Descriptive Statistics
§ Standard deviation (frequently noted as s or SD) is the average amount of
variability in a data. Tell researcher how far data lies from mean.

§ The larger standard deviation from mean (M), the more variable the data is

§ Manually you can used step below to determine the SD

1.List each score and find their mean.

2.Subtract the mean from each score to get the deviation from the mean.
3.Square each of these deviations.
4.Add up all of the squared deviations.
5.Divide the sum of the squared deviations by N – 1.
6.Find the square root of the number you found.

See tabulation of the steps on the next slide

 Descriptive Statistics
step 2 step 3

step 4
step 1
 Descriptive Statistics

§ Lastly, variance is the average of squared deviation from the mean reflecting
the degree of spread in the data set.

§ If the SD is 9.18, the s2 = 84.3

 Inferential statistics Statistics

v Are techniques that allow researcher to use samples and make

generalization that representing the population.

v To ensure this is achieve, the sample need to be accurately selected and

chosen using appropriate sampling strategy.

v Inferential statistics can be separated into several methods which include

estimation of parameters and testing statistical hypothesis
 Inferential statistics Statistics
v Most of the time, researcher cannot sample everyone in the population
due to lack of time, cost and manpower.

v Thus, descriptive statistics used to summarize the sample characteristics

but inferential statistics use to make guesses about the larger population
from the data collected through the samples and respondents

v Moreover, as the size of a sample always smaller than the size of a

population, due to some of the population isn’t be included by sample
data

v This lead to sampling error.

v Sampling error can be reduce using probability sampling methods.

 Inferential statistics Statistics
v Sampling error can be describe as difference
between
Point estimate
1. Parameters – true/actual population values • Single value estimate of
(population mean) parameters such as sample mean
2. Statistics – measured sample values (e.g. use to estimate population mean
sample mean).
Interval estimate
v In most cases, researcher didn’t know the actual • Give a range of values where
population parameter, thus, statistics is use to parameter is expected to lie
estimate these parameter taking sampling error • E.g. confidence interval
into account.

v Estimation on the population can be done through

two types of estimates.
 Inferential statistics Statistics

v Confidence interval uses the variability of statistic to come out with

parameter estimate

v Each confidence interval is associated with confidence level

v Let say, a 95% confidence interval set in research, if we repeat the

same study with a new sample in exactly the same way for 100 times,
we expect the estimate we get will lie within the specified range of
values for 95%.

v The higher confidence interval, the higher confidence level (e.g.99%).

v This is one of reason to determine the significance level in research for

example p<0.05 or p<0.01.
 Inferential statistics Statistics

v Hypothesis testing is a statistical analysis using inferential statistics

with main focus is to compare the populations or assess relationships
between variables.

v Hypotheses can be said as prediction in a study.

v Statistical test can be parametric or non-parametric with parametric

test consider more statistically powerful as they are more likely to
detect an outcome or effect.
 Inferential statistics Statistics

v Must be noted that each parametric tests have their assumptions and
when the data met the assumption, the test can be used to analyse the
findings

v If not, researcher must find appropriate test for his/her data

v Example of assumption in parametric test are

1. Samples must be normally distributed.

2. Sample size is large enough to represent specific population.
3. Variance, a measure of variability of each group compared are
similar.
 Inferential statistics Statistics

v However, in the case of data violates the assumptions, researcher can opt
to use non-parametric tests are more lenient as it do not rely on normally
distributed data.

v There are 3 forms of statistical tests

comparison correlation regression

 Inferential statistics Statistics

v Comparison test:

ü Assess whether there are difference in means, medians or

rankings of score between two or more groups.

ü Test need to be chosen while considering the assumptions of

parametric tests, number of samples and level of measurements
for variables.

ü Mean can be found for interval or ratio data.

ü Medians and rankings more suitable for ordinal data.

 Inferential statistics Statistics
v Example of Comparison test
 Inferential statistics Statistics

v Correlation test

ü Determine the extent of which two variables are associated

ü Term use usually ‘relationship’ between two variables

ü Most frequently use correlation test is Pearson’s r

ü Spearman’s r is appropriate for interval and ratio variables when

data is not normally distributed

ü Chi-square test of independence can be used for nominal

variables
 Inferential statistics Statistics

v Example of Correlation test

 Inferential statistics Statistics

v Regression test

Ø Demonstrate whether changes in predictor variables will cause

changes in the outcome variables

Ø Independent variable can be predictor variables and dependent

variable is the outcome

Ø Commonly, regression tests are parametric which requires data

which not normally distributed to be transform using mathematical
operation such as square root before performing regression
 Inferential statistics Statistics
v Example Regression test
 Inferential statistics Statistics
So what is normally distributed data

v If histogram being used, bell shape curve of the histogram showing

the distribution of the data
 Inferential statistics Statistics
So what is normally distributed data

v Data is not normal if it is skew or having kurtosis

v However, some of the easiest way of determine normal distribution is using

Kolmogorov-Smirnov test using SPSS.
 Inferential statistics Statistics
Example of Kolmogorov Smirnov test

• When u refer table below for reaction time trial 1, 2, 3,and 5 the value of p (asymptomatic
significance (2-tailed) were less than 0.05.
• In this case, it reject null hypothesis that the data is normally distributed, thus, concluded
these data were not normally distributed.
• Hence, instead of parametric test, non-parametric test will be used to analyse the data
• If the value for the data is p>0.05 like in reaction time trial 4, null hypothesis for normal
distribution is accepted, thus, data was normally distributed
 Inferential statistics Statistics
• In this case, student need to remember than the used of significance value in inferential test where
p<0.05 and confidence interval (CI) is 95% show significance mean difference, association,
relationship and etc.

• Similar this application may carry difference meaning in Kolmogorov- Smirnov test as p<0.05 show
that the distribution of data is not normal

• Student need to understand statistical test to be used in order to used them for analysis instead of
randomly select tests that they want.

• Moreover, each of parametric and non-parametric tests have assumptions that need to be met
before a test can be run.

• Refer to examples in the next slide

 Inferential statistics Statistics
Independence sample t-test (parametric test)
 Inferential statistics Statistics
Chi-Square Test of Independence Assumptions (non-parametric tests)
 Inferential statistics Statistics
Mann-Whitney U Test (Non-parametric test)
 Inferential statistics Statistics
If normal distribution of data cannot be achieve, below is the list of alternative
for parametric tests
THANK YOU
 REFERENCES
1. Prtha Bhandari. (2023). Descriptive Statistics: Definitions, Types, Examples. Scribbr.
Retrieved on 25 June 2023 from https://www.scribbr.com/statistics/descriptive-statistics/
2. Prtha Bhandari. (2023). Inferential statistics: An easy Introduction. Scribbr. Retrieved on 25
June 2023 from https://www.scribbr.com/statistics/inferential-statistics/
3. Laerd Statistics. (2018). Descriptive and Inferential Statistics. Retrieved on 24 June 2023
from https://statistics.laerd.com/statistical-guides/descriptive-inferential-statistics.php
4. Laerd Statistics. (2018). Chi-Square Test for Association using SPSS Statistics. Retrieved
on 24 June 2023 from https://statistics.laerd.com/spss-tutorials/chi-square-test-for-
association-using-spss-statistics.php
5. Laerd Statistics. (2018). Independent t-test using SPSS Statistics. Retrieved on 24 June
2023 from https://statistics.laerd.com/spss-tutorials/independent-t-test-using-spss-
statistics.php
6. Laerd Statistics. (2018). Mann-Whitney U Test using SPSS Statistics. Retrieved on 24
June 2023 from https://statistics.laerd.com/spss-tutorials/mann-whitney-u-test-using-spss-
statistics.php