5 NEW TRENDS IN SOCIAL SCIENCE RESEARCH
https://1558e42a91b2843f191224a9ed49c69c.safeframe.googlesyndication.com/safeframe/1-0-40/html/container.htmlPresentationPDF Available
Emerging trends in social science research
- December 2015
- Conference: Presentation at the Interactive Session with UG and PG students
- At: PES University, Bangalore South Campus, Electronic City, Bangalore 560100, India,
- Affiliation: Baze University Abuja Nigeria
Authors:
Download citationCopy link
Abstract
This presentation tried to revisit some of the evolving areas of concerns in contemporary social science research. Deficient knowledge of data and various typologies and how each system should be handled differently in data analytics, bias and sampling errors, fixation on computer outputs – wrong interpretation of analytical results, e.g. hypothesis test results, and inadequate training of enumerators or pilot studies are among areas of increasing concern in research practice. Increasing emphasis on interdisciplinary / collaborated research, technology-aided empirical research, relatively shorter duration of research in a ‘big data’ era, as well as quality and integrity of research reports, among others, have attracted increased attention in contemporary social science research. All these dynamics should drive the experienced practitioner and the budding researcher towards sustained production of more focused, quality research for shared well-being, prosperity, and progress.
Discover the world’s research
- 25+ million members
- 160+ million publication pages
- 2.3+ billion citations
Public Full-text 1
Content uploaded by Stephen Aro-Gordon
Author content
Content may be subject to copyright.
Emerging trends in
social science research
Presentation at the Interactive Session with UG and PG students held at PES University, Bangalore
South Campus, Electronic City, Bangalore 560100, India, on December 21, 2015
Stephen ARO-GORDON, Ph.D.
Baze University Abuja, Nigeria
Department of Financial Mathematics
Faculty of Computing and Applied Sciences
Email: stephen.aro-gordon@bazeuniversity.edu.ng/; getyouthsworking@yahoo.com
1
December 21, 2015 @ PES University, Bangalore, India
Objectives…
1. To revisit some general areas of concerns in
contemporary social science research.
2. To review some of the emerging shifts in
social science research especially in
emerging market economies including some
of the noted implications for the benefit of
future researchers in the field.
December 21, 2015 @ PES University, Bangalore, India
Resources…
1. Research Methodology: Methods and Techniques. Third Edition. By
Kothari, C. R., & Garg, G (2014). New Delhi: New Age International (P)
Limited, Publishers.
2. Research Methodology in the Social and Management Sciences by U.
G. Moti, Ed. (2005). Kaduna: Joyce Graphic Printers & Publishers
Company.
3. Discovering Facts Using SPSS by Andy Field
4. Building Financial Models with Microsoft Excel: A Guide for Business
Professionals, 2nd edn by S. Proctor (2009). Hoboken, NJ: Wiley.
5. Introduction to Linear Regression Analysis, 3rd edition, by Montgomery,
Peck, & Vining (2003). Wiley India.
6. Introductory Econometrics for Finance, 2nd edn. by Chris Brooks (2008).
Cambridge University Press.
December 21, 2015 @ PES University, Bangalore, India
Context…
What really
is ‘science’?
December 21, 2015 @ PES University, Bangalore, India
Understanding what ‘science’
is all about….(Encarta Dictionary, 2009)
1. Study of physical world: The study of the physical
and natural world and phenomena, especially by
using systematic observation and experiment.
Branch of science: a particular area of study or
knowledge of the physical world.
2. Systematic body of knowledge: Asystematically
organized body of knowledge about a particular
subject. Something studied or performed
methodically: An activity that is the object of careful
study or that is carried out according to a developed
method.
December 21, 2015 @ PES University, Bangalore, India
Science as a body of knowledge…
1.Social sciences
2.Physical sciences or natural
sciences
3.Life sciences
4.Management sciences, etc.
December 21, 2015 @ PES University, Bangalore, India
Exploits of physical science research
on CNN this morning! (21/12/2015)
December 21, 2015 @ PES University, Bangalore, India
Social science:
Science of society…
Social science is a key
discipline dealing with aspects of
the society –relationships
among groups of people within a
community.
Established thinkers, notably,
Karl Marx, Weber, Georg
Simmel, and Emile Durkheim
Contemporary influencers –Karl
Popper and Talcott Parsons
Typically –Economics, Political
Science, Public Administration,
International Relations, Geography,
Demography, Sociology,
In wider context –Anthropology,
Archeology, Psychology, History, Law,
Criminology, Linguistics, Education,
Communication Arts, Business /
Management Studies, Development
Studies, Environmental Studies,
Library Science, Marketing
Boundary is blurring –Sociobiology,
Neuropsychology, Behavioural
Finance, Bioeconomics, Social
Medicine
8
December 21, 2015 @ PES University, Bangalore, India
What ‘research’ involves…
Research = “to go about seeking” or ‘search’ (French “recherché”)
Research is basically a creative work conducted in a systematic manner
to add to stock of knowledge and help in devising new applications that
would make the society a better place.
Research has also been defined variously as:
Data / information / facts collection for advancement of knowledge
A process…
A studious inquiry or examination…
Empirical research is not derived from application of logic; rather, it is
based on:
Observation (sensory)
Experiment
Practical experiment
December 21, 2015 @ PES University, Bangalore, India
Revisiting the purpose of research –
The BIG PICTURE
Well-being of the society…
The current massive data availability –the
Internet age / smart phones
Rapid and monumental changes and
implications on the future of the professions –
e.g. technology ‘disruptions’ rendering
traditional approaches antiquated, opaque,
and unaffordable (Susskind &Susskind, 2015)
December 21, 2015 @ PES University, Bangalore, India
The question of
‘methodology’ and ‘methods’
December 21, 2015 @ PES University, Bangalore, India
The recurring question of
‘methodology’ and ‘methods’
Methodology is the general research strategy –
outlining the way in which you want to undertake the
research, the methods you will use, + the theoretical
justification for using the methods
‘Methodology’ is not synonym for method…
Methods captures the means or modes for
collecting data (publication research, interviews,
surveys, etc.) and in some cases, how you want to
analyze, compute or infer from the result.
December 21, 2015 @ PES University, Bangalore, India
Research methods
December 21, 2015 @ PES University, Bangalore, India
Exploratory research
-Helping to identify and define research problems
Constructive research
– Testing theories and proposing solutions
Empirical Research
-Testing the feasibility of solution using real-world data
Two major types of
empirical social research design (Gorard, 2013)
December 21, 2015 @ PES University, Bangalore, India
Qualitative research
-Understanding human behaviour
Philosophy: Social constructionism
Quantitative research
-Uses numerical data and statistical methods
Philosophy: Positivism
Formal social research process…
1. Identification of research
problem –the research gap
that you intend to narrow
2. Literature review
3. Specify purpose of your
research
4. Determine the specific
research questions
5. Formulate your hypotheses
–a testable prediction which
designates the relationship
between two or more
variables.
6. Select a suitable methodology
7. Data collection
8. Analyze and interpret the data
*9. Report and assess the
research results (was the Null
hypothesis validated or
rejected?)
*10. Communicate the
research findings,
recommendations and scope
for further research.
15
December 21, 2015 @ PES University, Bangalore, India
Wider variety of data sets…
Fundamental economics
Real estate
Human resources
management
Accounting
Advertising
Agriculture
Banking & Finance
Business
Finance & investments
Marketing
Opinion polls
Transportation
Sports
Life sciences
Physics and
engineering, etc.
16
December 21, 2015 @ PES University, Bangalore, India
Errors in data analytics
Areas to watch out for
December 21, 2015 @ PES University, Bangalore, India
Statistics revisited…
means or end?
Statistics is the scientific method
of collection, classification,
representation, analysis and
interpretation of numerical data
with a view to making
meaningful inferences on the
objects on which the data were
collected.
Statistics answers questions
using data –not numbers only;
we also use pictures, graphs,
tables, etc.
Statistical models are used to
simplify reality and help us to
answer questions.
Capacity to make
desirable, informed,
result-oriented
decision.
one thing is clear: we
use statistical
processes to serve as
guide towards making
qualitative decisions.
18
December 21, 2015 @ PES University, Bangalore, India
Are ‘numbers’ the end?
“Far too many reviews are dominated by
dry discussions of numbers… The review
should be a creative exercise, not a drill
where people regurgitate data.”
– L. Bossidy & R. Charan (2002) in Execution: The discipline of getting things done. New York: Crown Business, p. 208
December 21, 2015 @ PES University, Bangalore, India
Increased awareness of statistical errors
Continuation
Wrong methodology –A
good number of
unintentional mistakes
come from using the wrong
method to address
research questions,
thereby leading to
misinterpretation of results.
Choose and use correct
statistical methods for
every problem
Data organization: Avoid
error here by clearly
defining your variables
(observations) before you
record your data –what
does X1, X2, X3, … mean?
The name of a variable
should describe its
attributes.
20
December 21, 2015 @ PES University, Bangalore, India
Why exercising
good judgement is imperative…
Analytical tools are as
important to the modern
executive as pliers and
screwdrivers are necessary
to the auto mechanic.
Like a mechanic, the
analyst must know his
business well enough to
choose the proper
statistical tool to solve the
problem at hand.
21
December 21, 2015 @ PES University, Bangalore, India
Resolving several social questions
in organizational settings …
What are your competitors doing to
increase market share?
What are your competitors doing
differentlyin the area of customer
service?
How are your best-in-class
competitors handling cost, quality,
technology and HR?
Does your organization have the cost
structure (or capital structure) that
will allow it to compete profitably?
Assume your organization needs a
new organogram; what new sales
management skills will be required?
22
December 21, 2015 @ PES University, Bangalore, India
How sensible is this?
One research approach to every problem?
“I suppose it is tempting, if the
only tool you have is a hammer,
to treat every problem as if it
were a nail.”
– Abraham Maslow
December 21, 2015 @ PES University, Bangalore, India
Minimizing statistical errors
Further tips…
Understand the underlying
business theory / issues /
questions first before performing
statistical analysis –this is
pivotal to drafting meaningful
questionnaires.
The central point of statistics is
problem-solving –how are your
analyses helping your
organization or country to make
better decisions or policies?
Don’t carelessly round up data –
check to ensure that your data
add up –data credibility
The need to produce better
decisions and insights from the
massive data amount generated
in today’s world of business and
science.
Technology –Computers now
perform most of the calculations
that once dominated statistics
and related courses
Use Statistical packages /
software carefully –interpretation
of results must make sense to
you first before it can make sense
to your audience
24
December 21, 2015 @ PES University, Bangalore, India
Minimizing statistical errors
Emerging concerns…
“All models
are wrong,
but some
are useful.”
– George E. P. Best
25
December 21, 2015 @ PES University, Bangalore, India
Data normalization and standardization:
now a major issue…
Basically, to normalize data, traditionally this means to fit
the data within unity (1), so all data values will take on a
value of 0 to 1 (Ben Etzkom, 2012). Where applicable,
data should be normalized or standardized to bring all of
the variables into proportion with one another.
This is important so that the coefficients associated with
each variable will scale appropriately to adjust for the
disparity in the variable sizes, thereby reflecting meaningful
relative activity between each variable, i.e., a positive
coefficient will mean that the variable acts positively
towards the objective function, and vice versa.
December 21, 2015 @ PES University, Bangalore, India
On data normalization and standardization
Emergence of Post-Modern Portfolio Theory (PMPT)
Associated with software
entrepreneurs Brian M. Rom &
Kathleen Ferguson, PMPT is an
expanded risk-return paradigm
designed to address the major
practical limitations of CAPM/MPT –
the assumption that of a discrete,
normal (mean-variance) distribution
that may not accurately reflect
investment reality.
Thus, the lognormal distribution was
introduced as a more robust model for
the pattern of investment returns.
December 21, 2015 @ PES University, Bangalore, India
Avoiding data confusion / mishandling…
Qualitative data –Categorical –
Discrete – Nominal –Ordinal (Likert
scale) –we cannot easily measure or
count; e.g. gender, behaviour,
quality…
Performing purely quantitative
techniques such multiplication and
division on categorical data will yield
meaningless results.
Don’t put ordinal data in a pie chart!
Don’t carelessly round off data
particularly in pie charts.
Be careful with elaborate graphs
Clarity in knowledge of Mean, Mode,
& Median…
Quantitative data –data that we can
easily measure and count; e.g. age,
weight, height, sales, production
output, prices…Numerical –
Continuous –Interval –Ratio
Time series (trend analysis) –
changing values of a variable over
time / at different times.
Cross-sectional data –data that
measure attributes of different objects
at the same time –one-shot data.
Panel Data –Data collected on
various objects (individuals, countries,
etc.) for sequential periods –a
combination of time-series and cross-
sectional data.
28
December 21, 2015 @ PES University, Bangalore, India
Self-study…
Parametric
versus
non-parametric
In what
situations
should you use
chi-square?
29
December 21, 2015 @ PES University, Bangalore, India
Ensuring robust sampling
The essence of sampling is to obtain
maximum, accurate, and reliable
information about the universe with
the minimum sacrifice of money,
time, and energy.
Statistical regularity – randomness
Inertia of large numbers –more
inert (constant) than small ones
Persistence –same attributes as the
universe
Optimization –cost-effective &
efficiency
Validity –selected at random,
scientifically done
30
December 21, 2015 @ PES University, Bangalore, India
On Sample Size…
The issue is about the accuracy of your data,
perhaps, not necessarily your sample size.
The point is: How representative is your sample? –
(the young girl thought it was ice cream, but a ‘little’
taste was enough to disappoint her!!)
It is better to get a representative sample than a very
large sample [See Vijayalaskshmi & Sivapragasam
(2008) on Research Methods: Tips and Techniques]
December 21, 2015 @ PES University, Bangalore, India
Some tips on sample size…
(Vijayalaskshmi & Sivapragasam, 2008)
The more varied the data, the smaller the sample size needed
to attain the same level of accuracy –if we have many
variables in the study, the sample size should be relatively
small.
On the other hand, if we have only a few variables, we should
have a larger sample.
In cross tabulations, the desirable sample size depends on
the number of intended cells –a rough guide is to have at
least 20-30 study units per cell.
In the final analysis, the eventual sample size depends on a
compromise between what is desirable and what is feasible.
December 21, 2015 @ PES University, Bangalore, India
Dealing with sampling errors…
Unbiased errors –
accidental or arising in the
course of events or survey
Also called random
sampling error –only a part
of the universe is ultimately
observed.
Solution: Ensure that the
sample size is reasonably
large to neutralize this type
of error
33
Biased errors –
prejudice of the
analyst /
investigator
Note: Increasing
sample size will not
cure biased errors.
December 21, 2015 @ PES University, Bangalore, India
Clarity of statistical typology
Types of
Statistics
Descriptive
statistics
Inferential
Statistics
December 21, 2015 @ PES University, Bangalore, India
Inferential statistics…
What is ‘inferential’?
December 21, 2015 @ PES University, Bangalore, India
What is ‘inferential statistics’?
Inferential statistics –
statistical approach that helps
you to draw clear or evident
implications from analytical
results
December 21, 2015 @ PES University, Bangalore, India
Understanding data typology…
Types of
Data
Qualitative
data
Quantitative
data
Econometric
data
December 21, 2015 @ PES University, Bangalore, India
Still on data typology…
Types of
Data
Primary
data
Secondary
data
December 21, 2015 @ PES University, Bangalore, India
Choosing between primary
and secondary data… things to consider:
Nature and scope of your research –more data could be
needed to reinforce or modify current knowledge.
Availability of resources –quality primary survey could take a
lot of time and finance.
You can never have enough of data any way –secondary
data can provide supplementary evidence.
How credible and reliable is the source –data-collection
agency
Why insist on primary data when the data needed already
exist
December 21, 2015 @ PES University, Bangalore, India
Era of massive data…
S/NO
SOURCES Examples of data
1
ARCHIVES
General
Information
2
PRINT MEDIA
(Newspapers, daily, weekly or monthly
magazine)
Election result, Academic
Result, admission
list of
Schools, and other
news items
3
ELECTRONIC MEDIA
Census result, election result,
Radio, television,
internet/websites,
News items and
general
information.
4
MUSEUM (a cultural archive)
Information on antiquities,
cultural information.
5
NATIONAL BUREAU OF STATISTICS
PLANNING COUNCILS, CENTRAL
BANKS, WORLD BANK, IMF, ETC.
Wages and salaries of workers,
facts on national economy.
December 21, 2015 @ PES University, Bangalore, India
Secondary data:
Have they passed the test of scrutiny?
A major feature of secondary data is that it
may be fraught with bias, and may not be
so accurate or adequate, depending on
the purpose of the investigation.
Consequently, since the researcher is
typically not the original compiler of the
data, there is aneed for proper editing
and scrutiny of the secondary data in
order to make it appropriate for usage.
To do this, answers must be provided to
the pertinent questions listed here –(right
side of this slide).
When answers are satisfactorily provided
for each of these afore mentioned
questions satisfactorily; the secondary
data are said to have been scrutinized
and edited, thus made reliable for use by
researchers.
1. What is the type and
purpose of the institution
which the data emanated?
2. Are the data accurate and
adequate?
3. Are the data biased?
4. In what types of units are
the data expressed?
5. Are the data related to the
problem under study?
41
December 21, 2015 @ PES University, Bangalore, India
‘Data’ vs. ‘information’ – any real difference?
Not all pieces of information gathered about a
particular object can be referred to as statistical
data. Why?
*****Data is factual information often in the form of
facts or figures obtained from experiments or
surveys, and used as a basis for making
calculations or drawing conclusions. (Encarta Dictionary, 2009)
December 21, 2015 @ PES University, Bangalore, India
Revisiting…
Methods of data collection
December 21, 2015 @ PES University, Bangalore, India
On methods of data collection…
integrity & finance issues…
1.Direct Personal
Observation – FGDs
2.Personal interviewing –
directly or indirectly
(due to language
barrier) – robust
dialogue among
business leaders,
executives, policy
makers, customers, etc.
3. Data collected through
experimentation…
4. Data collected through
questionnaires…
reliability issue –how
truthful are the
respondents –how
truthful are the
enumerators, etc.
44
December 21, 2015 @ PES University, Bangalore, India
Data collection:
Problems and type of errors that may occur …
BIAS: The enumerator and
respondents may colour their
information with ethnic or tribal
sentiments.
SAMPLING: Sample may
be an incorrect representation of
the population.
ENUMERATOR: The
questionnaire may not be
properly administered, thus
required data may not be
collected as it ought to be
Deficiency of
Questionnaires
Reluctance on the
part of informant
45
December 21, 2015 @ PES University, Bangalore, India
Variables – revisited …
Types of
Variable
Discrete
variables Continuous
variables
December 21, 2015 @ PES University, Bangalore, India
On discrete & continuous variables
Discrete Variable
These are quantitative
properties of objects that
can be expressed in terms
of whole numbers only.
For example, the number of
a particular book sold by a
bookseller can only be
counted whole numbers:
10,12, 14 etc.
Continuous Variable
•These are quantitative
properties of objects that
are measured on a
continuous scale rather
than counting.
•For example, the heights of
school pupils are measured
in feet to the nearest
centimeters 1.89 metres
47
December 21, 2015 @ PES University, Bangalore, India
Why you need to classify data properly
•To ensure that data so
collected by any of the
data collection methods
be well-organized
thereafter.
•To improve our
understanding of
problems and helps us
to take wise decisions.
•To eliminate
unnecessary details.
•To bring out clearly
points of similarity and
dissimilarity.
•To enable us to form
clear mental picture of
objects.
•To enable us to make
comparison and draw
meaningful inferences.
48
December 21, 2015 @ PES University, Bangalore, India
How to make meaningful data
presentation…
December 21, 2015 @ PES University, Bangalore, India
Presenting statistical data
The simplest and
perhaps the most
important way of
presenting data is by
means of diagrammatic
form, design, or chart,
not only to catch the
eyes, but also to
convey the information
easily.
Pie chart
Bar chart
Simple bar chart
Component bar chart
Multiple bar chart
Histogram
Frequency polygon
50
December 21, 2015 @ PES University, Bangalore, India
Handling ‘normal distribution’…
December 21, 2015 @ PES University, Bangalore, India
Handling correlation analysis
December 21, 2015 @ PES University, Bangalore, India
Beta – a key application area in social
sciences, particularly the business world
Betas vary significantly across
industries.
Hence there are differences in
business risks across industries
(Chandra, 2008)
December 21, 2015 @ PES University, Bangalore, India
Interpretation of correlation coefficient:
Summary
S/N
o
r Interpretation
1
= 1
Perfect positive relationship
2
= -1
Perfect negative relationship
3
= 0
No relationship
4
= 0.1 to 0.4
Positive and weak relationship
5
= 0.5 to 0.9
Positive and strong relationship
6
= -0.4 to -0.1
Negative and weak relationship
7
= -0.9 to 0.5
Negative and strong relationship
December 21, 2015 @ PES University, Bangalore, India
Dealing with regression analytics…
December 21, 2015 @ PES University, Bangalore, India
Note the essence of
regression analytics (Montgomery et al, 2003)…
Modeling a complex system
Data description
Prediction & estimation
Parameter estimation & control
Model validation
oNote: There is no ‘best’ regression estimation
(Montgomery et al, 2003, p. 291)
December 21, 2015 @ PES University, Bangalore, India
Recap:
Interpretation of goodness of goodness of fit
S/No
R2Interpretation
1
= 1 Perfect fit
2
= 0 No fit
3
= 0.1 to 0.4
Poor fit
4
= 0.5 to 0.9
Good fit
STA204 STATISTICS
December 21, 2015 @ PES University, Bangalore, India
Recall interpretation of correlation coefficient:
S/No
r Interpretation
1
= 1
Perfect positive relationship
2
= -1
Perfect negative relationship
3
= 0
No relationship
4
= 0.1 to 0.4
Positive and weak relationship
5
= 0.5 to 0.9
Positive and strong relationship
6
= -0.4 to -0.1
Negative and weak relationship
7
= -0.9 to 0.5
Negative and strong relationship
December 21, 2015 @ PES University, Bangalore, India
Recall:
Interpretation of ‘goodness of fit’:
S/No
R2Interpretation
1
= 1 Perfect fit
2
= 0 No fit
3
= 0.1 to 0.4
Poor fit
4
= 0.5 to 0.9 Good fit
December 21, 2015 @ PES University, Bangalore, India
Revisiting the R-squared metric…
R-Squared is seen as a non-decreasing measure meaning
that even as more explanatory variables enter the model, the
R-squared value will keep increasing thereby failing to
account for the loss of degree of freedom arising from
additional independent variables in the model.
This deficiency is addressed by adjusted R-squared which
gives a truer picture of the goodness of fit of a statistical
model.
R2or adjusted R2> 0.5 is thought to imply that the regression
model has a good fit and it is therefore okay for forecasting –
Really?!!
December 21, 2015 @ PES University, Bangalore, India
How reliable are social science research results these days?
‘Hypothesis testing’ revisited
December 21, 2015 @ PES University, Bangalore, India
Revisiting…
What “statistical significance” really means
When a statistic is ‘significant’, what this really means in
statistics is that you are very sure that the statistic is
reliable, not that the finding is important or that it has any
decision-making utility to your client or audience.
Significance is a statistical term that indicates how sure or
confident you are that a difference or relationship exists,
and how strong or weak it is.
Significant differences can be large or small depending on
your sample size.
In effect, what we are talking about is not the ordinary
sense of significance but ‘statistical significance’.
December 21, 2015 @ PES University, Bangalore, India
What ‘hypothesis’ is all about…
Recall: Hypothesis is simply a theory needing investigation:
a tentative explanation for a phenomenon, used as a basis
for further investigation
Hypothesis is therefore an assumption or a statement that is
assumed to be true for the sake of argument.
As analysts, we usually want to start our investigations with
basic assumptions or intelligent guess concerning the
behaviour of variables or issue under study.
December 21, 2015 @ PES University, Bangalore, India
That market ‘assumption’ –
test it before launching that product!
Hypothesis testing –testing the validity of your
results / claims, scientific verification or validation to
confirm whether the assumptions made about the
subject variables are true or false.
Types of hypothesis testing – what do you
understand by
One-tailed test?
Two-tailed test?
December 21, 2015 @ PES University, Bangalore, India
On whether to use a one-tailed or
two-tailed test of significance?
December 21, 2015 @ PES University, Bangalore, India
Whether to use a one-tailed or
two-tailed test of significance?
The answer largely depends
on your hypothesis.
It is generally safest to use
two-tailed tests
[https://www.statpac.com/surveys/statistical-significance.htm –29/11/2015]
December 21, 2015 @ PES University, Bangalore, India
Hypothesis Testing (HT):
Revisiting its essence…
HT is perhaps an area in which statistics finds its
greatest applications –testing validity of claims.
Testing is needed in practically every field of
human activity, particularly business and
management research because new ideas and
products should be properly evaluated before they
are launched into the market place.
December 21, 2015 @ PES University, Bangalore, India
Hypothesis testing:
continuation…
Hypothesis tentatively explains an observation
that can be tested (i.e. proved or disproved) by
further investigation.
Figuring out the solution to the problem, i.e.
“hypothesizing”, before you start will help build a
roadmap for approaching the problem
You can express hypothesis as possible root
causes of the problem
Breaking down the problem into key drivers (root
causes) or aspects can help formulate hypothesis.
December 21, 2015 @ PES University, Bangalore, India
Four key attributes of a good
hypothesis
1. Simplicity, clarity, and precision –not
vague
2. Verifiable –testable
3. It should have clear nexus with the
research problem, the research
questions, and the researcher’s
specified objectives.
4. It should have clear nexus with the
relevant theory – it shouldn’t be a wild
guess.
December 21, 2015 @ PES University, Bangalore, India
Formulating your hypothesis…
Null Hypothesis: Ho
Ho is set up by the
investigator with the
intention of being
rejected based on
the available
statistical evidence.
Alternative Hypothesis: H1
H1is the hypothesis
accepted by the
investigator after the
Ho has been
rejected.
70
December 21, 2015 @ PES University, Bangalore, India
Type I and Type II decision errors
Type I error –
alpha (?)error:
True Hois rejected
False H1is
accepted
Type II error –beta
(?)error:
True H1is rejected
False Hois
accepted
Why is Type II error
usually emphasized
in statistical
literature?
71
December 21, 2015 @ PES University, Bangalore, India
Traditional testing statistics…
t-test statistics (the student t-test) –
small sample test, i.e. n< 30
Z-test statistics (normal test) –
large sample test i.e. n≥30
F-test statistics (joint test)
Chi-square distribution X2
December 21, 2015 @ PES University, Bangalore, India
‘Sacrosanct’ confidence levels…
Confidence level –the extent of confidence or certainty that the
investigator has; conventionally (2-tailed):
ZT= Z?= 1% = 2.58
ZT= Z?= 5% = 1.96
ZT= Z?= 10% = 1.65
Decision Rule:
Calculated test statistic > tabular (critical) value: Accept H1 and reject Ho
i.e. the parameter testes is statistically significant
Calculated test statistic < tabular value: Accept Hoand reject H1
i.e. the parameter testes is statistically insignificant
December 21, 2015 @ PES University, Bangalore, India
Statistical testing:
On simpler (less confusing!) approaches…
The p-value is the probability that the test statistic (z*, t*, x*,
F*) will be exceed, and thus pis called the observed level of
significance, in contrast to the ?-value which is a priori-level
of significance.
The default value of ?= 0.05, and the relationship between P
and ?is as stated below:
i. If p≥ ?, do not reject Ho
ii. If p< ?, reject Ho
74
December 21, 2015 @ PES University, Bangalore, India
Statistical testing:
On simpler (less confusing!) approaches…
In essence, using a table is not necessary when you have the
exact probability for a statistic.
Your suitable computer applications can calculate exact
probabilities for most test statistics. If you have an exact
probability output from computer software, you simply
compare it to your critical alpha level.
If the exact probability is less than the critical alpha level,
your finding is significant.
if the exact probability is greater than your critical alpha level,
your finding is not significant.
75
December 21, 2015 @ PES University, Bangalore, India
Type I and Type II decision errors…
Pick hypothesis before looking at
the data to avoid bias –apriori
expectation
Don’t confuse statistical importance
with substantive importance.
The p-value is not necessarily the
probability that the null hypothesis
(Ho) is true; the p-value already
assumes that Ho is true.
Rather, it is the probability of
rejecting Ho incorrectly on the
basis of your results that is
displayed in sample assessment.
You have learnt about how to
interpret the adequacy of statistical
models, such as beta coefficients, R-
squared, adjusted R-squared, finding
the critical value of Fstatistic and
verifying the Fcalculated value, but
ensure that you able to summarize
your results in everyday language –
less jargons, please!
76
December 21, 2015 @ PES University, Bangalore, India
Hypothesis testing
The standard procedure revisited…
I
Formulate
Hypotheses
II
Select confidence
level
III
Select
estimator
IV
Select tail
type
December 21, 2015 @ PES University, Bangalore, India
Hypothesis testing
The standard procedure revisited…
V
Calculate test statistic
VI
Compare t-value to the
rejection region
VII
Make your
conclusions
December 21, 2015 @ PES University, Bangalore, India
Statistical testing:
Technology is trending, but…
The manual procedure for hypothesis testing can be tedious.
Happily, there are statistical tools available these days to solve the
problems with relative ease, once they have been carefully / properly
defined.
All you need to do is enter the given data and then issue the appropriate
commands.
Software supporting hypothesis tests: SPSS, MINITAB, EViews, MS
Excel, and so on.
When the results of manual solutions are compared to the computer
solutions, we see that they are identical.
But…
79
December 21, 2015 @ PES University, Bangalore, India
Om effective and efficient application of the computers…
some provisos…
Understand the methodology –underlying principles and
procedure.
Be familiar with the output generated by the computer
(you will see some examples of these later in this
presentation).
Be knowledgeable enough to interpret the computer
results meaningfully (back to the recurring imperative of
proper understanding of the related theory, the business
and the question at hand).
80
December 21, 2015 @ PES University, Bangalore, India
Statistical computing
The case of SPSS…
IBM SPSS –Statistical Package for
Social Sciences, is widely used
computer-based analysis.
What it does / can do:
Enables you to input your raw data
You can modify your inputted data
81
December 21, 2015 @ PES University, Bangalore, India
The case of SPSS
continuation …
With SPSS, you can carry out a wide range of
simple, statistical and multivariate analyses such as:
Calculating means
Cross-tabulation
Correlation analysis
Regression analysis
ANOVA
Discriminant analysis, factor analysis, principal
component analysis, etc.
82
December 21, 2015 @ PES University, Bangalore, India
Checking the effectiveness of your research tool:
The 4 Big Questions…
Reliability –how reproducible or repeatable is your
diagnosis /results by another (independent) researcher using
the same tool / data?
Validity –Is the measure really measuring what it claims to
measure –i.e. really serving the purpose for which it was
constructed?
Practicability –How feasible, in terms of ease of
administration, scoring, interpretation, economy of time, cost
and energy?
Objectivity –Will there be disagreement as to what the
correct answer / solution is, or, it ‘depends’…?! [If it’s
objective, then, there should be only one correct answer]
83
December 21, 2015 @ PES University, Bangalore, India
To sum up – Common errors in data analytics:
Some areas to watch out for
1. Fixation on numbers at the expense of
intelligent decision-making – deficient
knowledge of what data analysis or
statistics is all about.
2. Wrong statistical tools/methods –
wrong judgment
3. Foggy definition of variables –
incomprehensible questionnaires
4. Un-standardized disparity in variable
sizes.
5. Assumption of normal distribution viz-a-
viz non-normal reality.
6. Using un-scrutinized data, especially
secondary data (what Prof Toyin Falola, 2015 called
“propagandistic data”, in The Punch, 11/6/2015, p. 26)
7. Deficient knowledge of data and
various typologies and how each
system should be handled differently
in data analytics.
8. Bias and sampling errors.
9. Fixation on computer outputs –
wrong interpretation of analytical
results, e.g. hypothesis test results
10. Absence or inadequate training of
enumerators – absence of pilot tests.
84
December 21, 2015 @ PES University, Bangalore, India
Next Session…
Broad areas of emerging shifts in
social science research & some
implications for researchers
December 21, 2015 @ PES University, Bangalore, India
The four global forces
breaking all the trends (Dobbs et al, 2015)…
1. Economic shift towards the emerging markets –Asia, Latin
America, Middle East, & Africa.
2. Technology –boundless information.
3. Demographics –falling fertility / migration / urbanization.
4. Trade and finance –increased global inter-connectedness –
Asia becoming the world’s largest trading region / rising
volume of trade between China and Africa (from $9bn in
2000 to $211bn in 2012).
December 21, 2015 @ PES University, Bangalore, India
Emerging shifts in
social research priorities (www.norrag.org)…
1. Qualitative research vs. quantitative
research: The flight to massive analytics /
digitization.
2. Increasing recognition / value attached to
research for knowledge advancement and
robust policy-making –[e.g. ICSSR
(India), NISER (Nigeria)]
3. Moving from basic research to
applied/empirical research
4. Does blind application of new
methodological techniques to
observational dataset represent progress in
applied research?
5. Increasing involvement in research
management, no longer just research
production.
6. Mixed-method research getting common –
qualitative + quantitative / primary data +
secondary data
7. Phenomena of ‘big data’.
8. Policy-relevant research as opposed to
‘abstract’ research / ‘academic’versus
‘practice-based’ research.
9. Specificity: Shift to location/region or
sector-specific –case studies
10. Duration: Shift to shorter duration; 3-6
months if not shorter
11. Increased interdisciplinary research.
12. Research professionalization: Private-
sector involvement –research-consultants
–data collection & compilation, not just
data analysis.
13. Increased emphasis on quality of
reports (with beautiful graphics, printing,
binding and all!)
14. Terminology change – ‘research study’
or ‘research report’?
87
December 21, 2015 @ PES University, Bangalore, India
Other related issues…
[see Kara, 2012 in Research and evaluation for busy practitioners: A time saving
guide. Bristol: The Policy Press)]
1. Anti-methodology / anti-methodological hegemony /imperialism –saying
‘No’ to one-cap-fits-all scientific methodology, or imposition of a particular
methodology (see also point 5 below).
2. Anti-secondary data –fixation on so-called ‘primary data’, but you can use
secondary data whenever justified
3. Quality of teaching vs. Quality of teaching
4. The issue of ‘original research ‘- research that produces completely new
knowledge; i.e. not based on existing knowledge in a new form.
5. Domination (via funding + publications + paper selection) of research ‘prowess’ /
gatekeeping by rich countries.
6. Academic integrity –Honesty, trust, fairness, respect & responsibility [Dr. (Mrs.)
A. T. Alabi, 2005 , in her ‘Ensuring academic integrity in research and publications’].
–role of project supervisors in ensuring this is being stressed.
7. Accounting properly for stochastic processes in time-series data.
December 21, 2015 @ PES University, Bangalore, India
Developments in
analytical tools and dimensions…
1.SPSS, MS Excel, EViews, etc.
2.Model building – Correlation – Regression
3.Factor Analysis –Multivariate Data Analysis –
Multiple Responses –Principal Component Analysis
–Dimension Reduction
4.Defining variables
5.Correspondence Analysis
6.Discriminant Analysis
7.Cluster Analysis, etc.
December 21, 2015 @ PES University, Bangalore, India
A word about research funding…
Corporate Sources
Private Foundations
Government research
councils
December 21, 2015 @ PES University, Bangalore, India
Re-emergence of time series data!
In brief
December 21, 2015 @ PES University, Bangalore, India
The issue of ‘spurious correlation’…
“First impressions
can often lead to
wrong conclusions.”
-Our Daily Bread, 2015, p. 346.
December 21, 2015 @ PES University, Bangalore, India
‘Spurious Correlation’…
Correlation does not mean
causation
Consequently regression may be
also spurious and interpretation may
not be valid.
93
December 21, 2015 @ PES University, Bangalore, India
Understanding time-based data…
In time-based data, the change in a
variable is an important concept.
The change in a variable yt, also known as
its first difference, is given by Δyt= yt–yt-1.
Δytis the change in the value of the
variable y from period t- 1 to period t
94
December 21, 2015 @ PES University, Bangalore, India
Understanding time series
analysis…
Time-based data –showing the dynamic movement
of a phenomenon over a period of time, usually at
equal intervals.
Used in any domain of applied sciences and
engineering that involve temporal measurements.
Sequence of data points –successive movements
over time, e.g. ocean tides, closing stock prices on
Indian stock markets, etc.
December 21, 2015 @ PES College, Bangalore, India
Beyond analytics:
KIV important workplace skills…
Written & oral communication
Problem-solving abilities –analytical reasoning
Creative thinking
Interpersonal skills (team work)
Time management
General professionalism, personal effectiveness, and work ethics
December 21, 2015 @ PES University, Bangalore, India
A word about
research ‘recommendations’…
It’s fine to include ‘recommendations’ at the end
of your paper in applied research to show that
you understand the importance / implications of
your research.
However, let you’re the recommendations have
clear nexus with the text, specifically the
evidence / results that you have presented in the
paper.
Don’t add ‘recommendations’ just for the sake of
it!
December 21, 2015 @ PES University, Bangalore, India
Keywords
Descriptive Statistics, Inferential Statistics, Time Series,
Econometric Data, Bias, Sampling Errors, Spurious Correlation
& Regression, Time Series Data, Stationarity, Data Collection,
Data Analysis, Correlation, Regression, Hypothesis Testing,
Enumerators, Pilot Tests, Technology. SPSS, MS Excel,
EViews, Model Building, Factor Analysis, Multivariate Data
Analysis, Multiple Responses, Principal Component Analysis,
Dimension Reduction, Defining Variables, Correspondence
Analysis, Discriminant Analysis, Cluster Analysis.
December 21, 2015 @ PES University, Bangalore, India
Concluding remarks
This presentation tried to revisit some of the evolving areas of concerns
in contemporary social science research. Deficient knowledge of data
and various typologies and how each system should be handled
differently in data analytics, bias and sampling errors, fixation on
computer outputs – wrong interpretation of analytical results, e.g.
hypothesis test results, and inadequate training of enumerators or pilot
studies are among areas of increasing concern in research practice.
Increasing emphasis on interdisciplinary / collaborated research,
technology-aided empirical research, relatively shorter duration of
research in a ‘big data’ era, as well as quality and integrity of research
reports, among others, have attracted increased attention in
contemporary social science research. All these dynamics should drive
the experienced practitioner and the budding researcher to sustained
production of more focused, quality research for shared well-being,
prosperity, and progress.
December 21, 2015 @ PES University, Bangalore, India
100
Stephen ARO-GORDON, Ph.D.
Department of Financial Mathematics
Faculty of Computing & Applied Sciences
Baze University Abuja, NIGERIA
Email: stephen.aro-gordon@bazeuniversity.edu.ng /
getyouthsworking@yahoo.com
December 21, 2015 @ PES University, Bangalore, India
Caveat / feedback …
All the materials used in this presentation are solely for
educational purpose, advancement of knowledge and
improvement of educational practices to move the society
forward. The author will be pleased to make good any
omissions or rectify any mistakes brought to his attention at
the earliest possible opportunity.
1
0
1
101
December 21, 2015 @ PES University, Bangalore, India
Citations (7)
References (0)
… It is expected that the knowledge and expertise of social sciences will be able to solve challenges like poverty, climate change and the food crisis as and when required (Caillods & Jeanpierre, 2010).The contemporary issues in social science research like more emphasis on interdisciplinary/multidisciplinary/collaborated research, use of technology in empirical research, relatively shorter duration of research in a “big data” era, as well as quality and integrity of research reports etc. have been able to attract attention of researchers all over the world. All these dynamics should drive the experienced practitioner and the budding researcher towards sustained production of more focused, quality research for shared well-being, prosperity, and progress (Stephen, 2015). …
… Skinner (2011) extends in his article “The Age and Ideas of Adam Smith” the relevance among his peers and successors of economic history, philosophy, and the history of ideas in the eighteenth century, as well as to all those involved in the study of Adam Smith. Stephen (2015) attempts to revisit some of the evolving areas of concerns in contemporary social science research and reported that deficient knowledge of data and various typologies and how each system should be handled differently in data analytics, bias and sampling errors, fixation on computer outputs -wrong interpretation of analytical results, e.g. hypothesis test results, and inadequate training of enumerators or pilot studies are among areas of increasing concern in research practice. …
Emerging Research Trends in Social Sciences
Book
Full-text available
- Mar 2019
ViewShow abstract
… The contemporary issues in social science research like more emphasis on interdisciplinary/multidisciplinary/collaborated research, use of technology in empirical research, relatively shorter duration of research in a ‘big data’ era, as well as quality and integrity of research reports etc. have been able to attract attention of researchers all over the world. All these dynamics should drive the experienced practitioner and the budding researcher towards sustained production of more focused, quality research for shared well-being, prosperity, and progress (Stephen, 2015). Skinner (2011) extends in his article ‘The Age and Ideas of Adam Smith’ the relevance among his peers and successors of economic history, philosophy, and the history of ideas in the eighteenth century, as well as to all those involved in the study of Adam Smith. …
… Skinner (2011) extends in his article ‘The Age and Ideas of Adam Smith’ the relevance among his peers and successors of economic history, philosophy, and the history of ideas in the eighteenth century, as well as to all those involved in the study of Adam Smith. Stephen (2015) attempts to revisit some of the evolving areas of concerns in contemporary social science research and reported that deficient knowledge of data and various typologies and how each system should be handled differently in data analytics, bias and sampling errors, fixation on computer outputs -wrong interpretation of analytical results, e.g. hypothesis test results, and inadequate training of enumerators or pilot studies are among areas of increasing concern in research practice. …
Social Science Research: An Introduction
Chapter
- Mar 2019
International Journal of Innovative Knowledge Concepts
Research
Full-text available
- Jun 2023
Impact of E-Learning Environment on Student Performance Using Data Mining Classification Techniques
Article
Full-text available
- Aug 2022
ViewShow abstract
Modern Trends in Research Methodology
Article
Full-text available
- Feb 2019
ViewShow abstract
Translações no Facebook: a controvérsia ‘Galo Machista’!? nas páginas de torcidas organizadas
Article
- Apr 2017
ViewShow abstract
Translações no Facebook: a controvérsia ‘Galo Machista!?’ nas páginas de torcidas organizadas
Article
Full-text available
- Apr 2017
ViewShow abstract
Recommended publications
Article
Unverifiable Accomplishments and Publications on Applications for Gynecologic Oncology Fellowships
March 2012 · Obstetrics and Gynecology
Selection of physicians for fellowships in obstetrics and gynecology subspecialties has become increasingly competitive. The number and quality of research publications is an important factor in the selection process. We sought to estimate the incidence of unverifiable publications among gynecologic oncology fellowship applicants. We reviewed the applications to a single gynecologic oncology … [Show full abstract]
Article
Full-text available
A review of the process of knowledge transfer and use of evidence in reproductive and child health i…
August 2018 · Health Research Policy and Systems
Abstract Background The paper carries out a situational analysis to examine the production, dissemination and utilisation of reproductive and child health-related evidence to inform policy formulation in Ghana’s health sector. Methods The study used Wald’s model of knowledge production, transfer and utilisation as a conceptual model to collect relevant data via interviews and administration of … [Show full abstract]
Article
Full-text available
An Exploration of the Life Cycle of eScience Collaboratory Data
The success of eScience research depends not only upon effective collaboration between scientists and technologists but also upon the active involvement of information scientists. Archivists rarely receive scientific data until findings are published, by which time important information about their origins, context, and provenance may be lost. Research reported here addresses the lifecycles of … [Show full abstract]
Article
Full-text available
The relationships between person-environment fit and job-related variables
May 2019 · Korean Journal of Industrial and Organizational Psychology
The purpose of this study is to meta-analyze the relationship between person-environment fit(person- organization fit and person-job fit) and job-related variables(organizational commitment, job satisfaction and turnover intention), and to investigate whether the relationship is moderated by organizational typology(public organization and private organization). A total of 15,589 employees from 51 … [Show full abstract]
or
Company
Support
Business solutions
© 2008-2023 ResearchGate GmbH. All rights reserved.
We and our partners use cookies✕