Research Methodolog y Dr. Chowdhury Saleh Ahmed
Jan 15, 2016
Research Methodology
Dr. Chowdhury Saleh Ahmed
1. Meaning of Research2. Objectives of Research3. Motivation of Research4. Types of Research5. Research Approaches6. Significance of Research7. Research Methods and Research
Methodology8. Criteria of good Research9. Problems of research in countries
like Bangladesh
Lecture 1
Meaning of Research
Research means an objective and systematic search for pertinent information on a specific topic.
Example of Research:
• Carbon di oxide emission due to traffic jam• Causes of crimes by slum dwellers etc• Effectiveness of Educational incentive system for
controlling drop-outs
Cont.. • Research has to be an original contribution to the existing
stock of knowledge.
Thus Research involves:
• Enunciating the problem• Formulating the hypothesis• Collecting data/ facts/information• Analyzing the data/ facts/ information• Reaching at conclusions
Meaning of Research: Example
• Problem: CO2- Traffic Jam is a health hazard-
• Hypothesis – within tolerable limit• Collecting CO2 data through a sampling• procedure• Analyzing data whether within tolerable
limit • Making Conclusion
Objectives of Research
• To achieve new insights into a phenomenon
• To know about the existing phenomenon
• To know extent of a cause or effect variables
• To establish relationship between variables
Motivation of Research: What makes people undertake research
• Desiring solution to the problem/ hazard
• Desire to do a social / national / organizational service
• Inner satisfaction of a creative job
Types of ResearchDescriptive versus Analytical :
Descriptive research means describing state of affairs – Mostly involves primary data – ex-post research
Analytical research means carrying out analysis on a phenomenon – Mostly involves secondary data – ex-ante research- before contemplated change
Cont.
• Applied versus Fundamental Research
• Applied research aims to find a solution to an immediate problem facing a country/ society/ organization etc.
• Fundamental research is a basic research mainly concerned with generalizations
Cont.
• Quantitative versus Qualitative Research
```Quantitative research is based on the measurement of quantity or amount
```Example CO2 emission````Qualitative research is concerned with subjective
attitude/ quality/ attributes / desires / feelings/ etc```Example : Feeling about Dhaka city air pollution
( intolerable / Mildly intolerable / tolerable
Cont..
Conceptual versus Empirical
Conceptual research is about abstract ideas or theory
Attack from ghosts,
Empirical Research is data-based and subject to verification
Research Approaches
• According to types of research discussed:
• There can be two approaches to research:
• Quantitative approach:
• Qualitative approach
Cont….Research Approaches
• Quantitative approach can be further divided into 3 groups:
• Inferential approach – infer characteristics of a phenomenon through collection of data –example CO2 emission.
• Experimental approach – Researcher deliberately changes some variables to know the causal effect- example – effect of temperature rise on plant growth under open sky
Cont… Research Approaches
• Simulation approach – Researcher artificially construct an environment and collects data
• A greenhouse is constructed and impact of temperature rise on plant growth is measured
Significance of Research
“ Doubt is often better than overconfidence as doubt leads to enquiry and enquiry leads to invention”
“ Research inculcates scientific and inductive thinking that promotes development of logical thinking”
Cont… Research Approaches
• Research provides the basis of all government / international policies
• Research is used for solving various problems of Businesses, NGOs, Societies etc.
Research Methods and Research Methodology
Research methods refer to use of instruments for a research technique
Example:
Type of Research
Quantitative Field Research
Methods of Research
Questionnaire
Focussed Group Discussion
Technique of Research
Researcher uses open and close ended questions
Researcher selects a particular group for the detailed discussion
Cont… Research Methodology.
• Research methodology is a way to systemically solve the research problem
• Research methodology has a broader dimension than research methods
Cont.. Research Methodology
• Elements of Research Methodology are: Defining objective
• Reviewing Literature• Formulating Hypothesis• Designing sample• Collecting data• Analyzing of data• Arriving at conclusions
Criteria of good Research
• The objective of the research clearly defined• The research methodology used should be
described in sufficient detail to permit another researcher to repeat the research for further advancement
• The sampling design should be such as to yield least error
• The writing should be done with complete frankness – nothing which has bearing on the result should be hidden
Cont. Criteria of good Research.
• The validity and reliability of data as a well as calculations should be re-checked to avoid mistakes
• Conclusions should be confined to those justified by data and analysis
• Researcher should be a person of integrity
Problems encountered in countries like Bangladesh
• Insufficient interaction between Academics and Practitioners
• Lack of training on the part of researchers
• Fear that information provided to researchers may be used against them
• Secondary sources of data are not timely available
Way forward to overcome research problems in Bangladesh• Greater interaction between academic institutions
and practicing organizations/ government/ NGOs etc.
• More education and training on research methodology
• National Policy on Research mentioning that research data can not be used against the person
• Timely publication of secondary data and enforcement of Act on “Right to Information”
The End
Defining the Research Problem
Defining the Research problem is the first step of Research Methodology or Research process:
Defining research problem
Formulating Hypothesis
Designing sampling technique
Collecting Data
Analyzing Data
Report Writing
Lecture 2
Research problem arises only when:
•There must be group/individuals facing the problem to be researched.
•There must be some objectives to be achieved from the solution of the problems
•There may be alternative means of obtaining the objectives
•Researcher must have some doubts about the relative efficacy of the alternatives
Points to be observed in selecting a Research Problem
•Subject on which research has been done should
not be chosen
•Controversial issues should be avoided
•Narrow or too wide issues should be avoided
•Research problem selected should be feasible
within means available
•Researcher should have some background
information on the research problems
Techniques of defining Research problem
1. Statement of the Problem in a general way2. Deep understanding the nature of the
problem3. Surveying the available Literature4. Developing the idea further through
discussion 5. Finally, Rephrasing the Research Problem
Techniques Involved in Defining a Research Problem
Statement of the Problem
First of all the problem should be stated in a general way
Understanding the nature of the problem
The researcher should be thoroughly knowlegible in the subject
The researcher should first discuss the problem with those who first raised the issue/ problem.
The researcher than should discuss the issue with the resource persons excelling in the subject
Surveying the available Literature
All available research concerning the problem at hand must necessarily be surveyed and examined before formulating the research problem.
This means, the researcher must be well conversant with available reports, records and literature.
Developing the idea through discussion
Discussion concerning a problem often produces useful information.
People with rich experience are in a position to enlighten the researcher on different aspects of the proposal.
It helps sharpen the focus on specific aspects of the research.
Rephrasing the Research Problem
Finally the researcher must rephrase the research problem into a working proposition.
Once the nature of the problem has been clearly understood, literature has been reviewed, discussion over the problem has taken place, this rephrasing the research problem into analytical / operational terms become relatively easy.
Additional Points • Technical terms and phrases with special
meanings should be clearly defined for general readership.
• Basic assumptions relating to the research problem should be clearly stated.
• Aim or value of the research should be stated.• The suitability of the time period and the
source of data availability should be considered.
• The scope of investigation or the limits within which the problem will be studied need to be mentioned.
Example of a too broad /non-specific topic:
Why is labour productivity lower in Bangladesh compared to Vietnam • Vague in terms of which sector
• Vague in terms of time frame
• Non-analytical – labour productivity depends on certain factors -
Cont.• Rephrasing:
• Factors responsible for productivity differentials in Bangladesh and Vietnam’s RMG sectors between 2005-10.
The End
Research Design
• Research design is “decisions” regarding what, where, when, how much, by what means etc.
• It is management of conditions for collection of data, analysis of data and report preparation on the research problem.
Lecture 3
Meaning of Research design
1. What is the study about?2. Why the study is being undertaken?
3. Where will the study be carried out?4. Where can the required data be found?5. What will be the sample design 6. What period of time the study will include?
cont…
7. What type of data is required? 8. What techniques of data collection will be
used?
9. How many items will be observed? 10. How will the data be analyzed?
11. In what format, the report will be prepared within given time and budget?
Cont… Meaning of Research design
Components of Research Design
• From last two slides, it follows that Research Design has five components:
1. Problem formulation and objective
2. Sampling design: design which deals with method of selecting items to be observed in the given study. (3-6)
3. Observational design: design which relates to the conditions under which the observations are to be made on the selected items. (7- 8)
4. Statistical design: design that deals with how many items will be observed and how information collected will be analyzed. (9 - 10)
Cont..Components of Research Design
• Operational design: design which deals with the techniques by which procedures specified in the sampling, observational and statistical designs can be carried out. Within given cost and time (11)
Research design must have:
1. Statement of the problem and objectives 2. Sources of information to be collected
(Sampling designs) 3. Types of information to be collected
(Observational designs)4. Approach to be used for collecting and
analyzing data (Statistical designs)5. Estimates of time and cost for the research
(Operational designs)
In Summary, Research design must have
• Clear Statement of the Research Problem and the objectives of research
• Sampling design
• Observational design
• Statistical design
• Operational design
Important Concepts Relating to Research Design
Important Concepts Relating to Research Design
• Dependent and Independent Variables
• Extraneous variable
• Control variable
• Research hypothesis
• Testing significance of the result on Research hypothesis
Dependent and Independent variable• Variable – A concept/ entity that can take different
quantitative values is called a variable• Continuous vs. discrete variable • Example:
• Individual’s earning depends on his/her knowledge and skill
• Here knowledge and skill is independent variable• Individual’s earning is dependent variable
Extraneous variable
• Variables not related to the study but affect the dependent variable.
• Example - Measuring Dependency of Rice yield to fertilizer doze in different districts.
• But soil types of different district would affect yield- Soil types are extraneous variables.
Control Variable
• Control variables are used to overcome the effects of extraneous variables.
• Example: BRRI’s sub-stations in different districts have experimental stations with normal soil types and temperature – These are used as control fields.
Research Hypothesis
• When a prediction or a hypothesized relationship is tested by scientific methods, it is termed as research hypothesis.
• The opposite of Research hypothesis is known as null hypothesis.
Example of Research and Null Hypothesis
Research Hypothesis:
Paddy yield depends positively on fertilizer applied.
Null Hypothesis:
Paddy yield has no relationship at all with fertilizer applied
The End
Testing Significance of the Result
• Why done?
• Because a sample is taken rather than whole population, therefore there is a need to test significance or confidence on the result.
Lecture 4
Testing Significance on the Result
• How it is done?
• Sample is said to be replica of the population. Population’s
distribution
Sample’s distribution
2.5% 2.5%Mean Height in cm
Testing Significance on the Result
If sample mean is known and standard deviation is known, then 95% confidence interval can be calculated.
Stages of Research Design
• Problem formulation and objective framing
• Sampling design
• Observational design
• Statistical design
• Operational design
Formulating Research Problem and
Objectives to be attained
• Survey of Literature
• Discussion with persons affected by the problem, Resource persons
Sampling Design :Relevant Terminology
• Universe / Population –entire research area
• Census – survey of entire population
• Sample Survey – surveying a part of the population
• Sampling Frame - List of population from which sample will be drawn
Criteria of Sampling:
Aim should be to avoid Systematic bias.
Systematic Bias occurs when:
• Inappropriate sampling Frame• Defecting Measuring device• Non-respondents• Indeterminacy Principle• Natural bias
Sampling Errors
• Sampling Errors is the difference between the sample estimate and the true population parameter.
• The sampling error can be found by subtracting the value of a parameter from the value of a statistic.
• Example : Sample height – Population Height of individuals
• Sampling error depends on sampling design.
Different Types of Sample Design
• There are basically 2 types of Sampling:
• Probability sampling and Non-probability sampling:
• Probability sampling Random Sampling-
• EACH SAMPLING UNIT HAS EQUAL PROBABILITY
Non-probability sampling non-random sampling-
EACH SAMPLING UNIT HAS UNEQUAL PROBABILITY,
Unrestricted vs. restricted sampling
• When each sample element is drawn individually and directly from the population at large, then sample drawn is known as un-restricted sample.
• CHART SHOWING BASIC SAMPLING DESIGNS Representation Basis
Element Selection
Technique
Unrestricted
Restricted Sampling
(Researcher’s individual judgment involved)
Probability Sampling
Non-probability Sampling
Simple Random Sampling
Haphazard sampling/ Convenience sampling
Stratified Sampling
(Researcher’s individual judgment involved)
Purposive sampling
(Researcher’s individual judgment involved)
Probability Sampling
• Known also as Random / chance sampling
• Here every item of the universe has an equal chance of inclusion in the sample.
• All possible samples have equal chance of inclusion
Cont.. Probability Sampling
• Therefore sample has the same characteristics of the population- replica of the population.
• Errors of estimation or significance of the results can be measured.
The End
Procedure of selecting a random sample
• Suppose we have to randomly select 3 people from the class of 60 students.
• ------------------------------• 1-60 numbers corresponding to the students
are written in 60 pieces of paper • These are folded so that numbers are not
seen• The paper slips are thoroughly mixed
Lecture 5
Procedure of selecting a random sample
• Then 1 piece of paper is selected without seeing the numbers written. The number is returned to the pool.
• Then 2nd piece of paper is chosen and then the 3rd paper. If same no. is chosen, process is repeated.
• Suppose 34, 03, 58 numbers/students are chosen
• Each number and each possible sample (such as 01, 60, 45, or 43, 06,55 has equal chance of selection) (1/60 x 1/60 x 1/60)
Random Table
47 91 82 28 81 95 70 89 73 48
10 4 41 40 86 27 46 80 20 58
24 34 43 50 12 33 90 3 96 38
97 11 63 21 55 99 68 87 29 78
52 31 71 1 39 62 49 16 88 66
75 15 22 74 59 26 85 100 25 76
35 2 93 14 92 13 36 54 72 37
23 83 32 64 79 45 5 61 17 77
8 44 7 19 30 60 94 56 98 67
53 9 51 69 42 84 18 65 6 57
Systematic sample: mix of random and non-random
sampling• Supposing, we have to chose 4 students
from 100 students.
• First a number from the random table is selected
• Then 25 is added to the number to select the next number. If total number exceeds 100 then move to the beginning.
Example
18 43 68 93
80 05 30 55
Initial chosen number and then 25 is added successively.
Advantage and disadvantage of Systematic Sampling
Advantage• Spread over evenly over the entire population
compared to random sample• Easier and less costlier methodDisadvantage
If any systematic bias on the ith item (e.g.., 25th item), it persists.
Otherwise, systematic sample is considered equivalent to random sample,
Random Stratified Sampling
• Stratified sampling is used when the population is not homogenous.
• Under Stratified sampling, more pertinent information about the different homogenous stratum is obtained and
• therefore better information for the whole population is obtained.
Cont…• However, various strata are to be formed in a
way so as to ensure that elements are most homogenous within strata and most heterogeneous between different strata.
• Otherwise no advantage from stratification
Cont. Random Stratified Sampling
• Here, Strata are purposively formed and involves past experience and personal judgment.
• Once strata are selected, selection of unit must be done on random basis
• For better result, sampling units taken from each strata should be proportion to the size of the strata.
Cluster sampling
• If total area of interest is large enough, area can be divided into a number of non-overlapping areas and then to select a number of smaller areas called clusters.
• The samples are units in these small areas or clusters.
Example
• Salinity prone areas subdivided into different unions can be clusters.
• Farmers belonging to the union can be sampling units who could be interviewed.
Multi stage sampling- Two stage Sampling
• Suppose we want to measure efficiency of Nationalized Commercial Banks (NCBs) of Bangladesh:
• First stage is to randomly select some 2/3 divisions.
• Second we can select some districts randomly and interview all bank managers of nationalized banks in these districts.
Cont……………..
• This is two stage sampling.
• Divisions
• Districts
• Sampling units are bank branches
Multi stage sampling -Three stage Sampling
• If instead of interviewing all bank managers in the districts, we go down one step and randomly select some towns in these districts. We then interview all bank managers of the towns.
Cont….
• Then this is a three-stage sampling.
• Divisions
• Districts
• Towns
Sampling units are bank branches
Multi-stage Sampling with probability proportional to size
• Here Probability of inclusion of a cluster/ town is proportional to its size (in terms of bank branches).
Example: Taking a sample of 10 branches from a total of 500 branches spread over 15
towns so that probability of selecting a town is proportional to size of the town (measured in
terms of branches)• Number of bank branches in 15 towns are as
follows:
1
35
2
17
3
10
4
32
5
70
6
28
7
26
8
19
9
26
10
66
11
37
12
44
13
33
14
29
15
28
• Table1
35
35
10
2
17
52
3
10
62
60
4
32
94
5
70
164
110
160
6
28
192
7
26
218
210
8
19
237
9
26
263
260
10
66
329
310
11
37
366
360
12
44
410
410
13
33
443
14
29
472
460
15
28
500
Cumulative branchesSequential number of branch selected starting with a random number 10
Comparison with Alternative sample design: Simple Random Sample
Randomly choosing 10 branches from 500 branches
No assurance of rightly covering all divisions, districts, and towns.
Bigger towns might have lower samples.
Complex/ Stratified sample is more justified.
Non-Probability Sampling
• Non probability sampling is one where there is no guarantee that sampling element has equal probability of being selected.
• It is a kind of deliberate sampling.
• It is also known as purposive sampling.• Example: For examining extreme poverty level of
draught prone area, some unions of Rangpur is purposively chosen and compared with a normal union of another district.
Limitation of Non-probability/ Purposive Sampling
• Researcher can purposively choose an area which best suits his point of view.
• Element of human bias is always there
• Used in small scale research.
• Rarely used in large scale research
Conclusion
• Direct/ Simple Random sampling should be attempted as it has lowest bias and more importantly significance of the result/ confidence on the result can be estimated.
• If known characteristic of the population is known before hand and if random sampling is difficult and researcher has no bias towards a particular result then purposive sampling is suitable.
The End
Group 1Assignment: Take a sample of 10 branches from a total of 500 branches spread
over 15 towns so that probability of selecting a town is proportional to size of the town
(measured in terms of branches)• Number of bank branches in 15 towns are as
follows:
1
18
2
17
3
10
4
32
5
87
6
28
7
26
8
19
9
26
10
66
11
37
12
44
13
33
14
29
15
28
Group 2 Assignment: Take a sample of 10 branches from a total of 500 branches spread
over 15 towns so that probability of selecting a town is proportional to size of the town
(measured in terms of branches)• Number of bank branches in 15 towns are as
follows:
1
8
2
17
3
10
4
32
5
87
6
28
7
26
8
19
9
26
10
76
11
37
12
44
13
33
14
29
15
28
Group 3 Assignment: Take a sample of 10 branches from a total of 500 branches spread
over 15 towns so that probability of selecting a town is proportional to size of the town
(measured in terms of branches)• Number of bank branches in 15 towns are as
follows:
1
8
2
67
3
10
4
32
5
87
6
28
7
26
8
19
9
26
10
26
11
37
12
44
13
33
14
29
15
28
Group 4 Assignment: Take a sample of 10 branches from a total of 500 branches spread
over 15 towns so that probability of selecting a town is proportional to size of the town
(measured in terms of branches)• Number of bank branches in 15 towns are as
follows:
1
18
2
67
3
10
4
32
5
97
6
28
7
26
8
19
9
26
10
26
11
37
12
44
13
33
14
29
15
28
Group 5 Assignment: Take a sample of 10 branches from a total of 500 branches spread
over 15 towns so that probability of selecting a town is proportional to size of the town
(measured in terms of branches)• Number of bank branches in 15 towns are as
follows:
1
38
2
67
3
10
4
32
5
97
6
28
7
26
8
19
9
26
10
26
11
37
12
24
13
33
14
29
15
28
Group 6 Assignment: Take a sample of 10 branches from a total of 500 branches spread
over 15 towns so that probability of selecting a town is proportional to size of the town
(measured in terms of branches)• Number of bank branches in 15 towns are as
follows:
1
38
2
67
3
90
4
32
5
17
6
28
7
26
8
19
9
26
10
26
11
37
12
24
13
33
14
29
15
28
Group 7 Assignment: Take a sample of 10 branches from a total of 500 branches spread
over 15 towns so that probability of selecting a town is proportional to size of the town
(measured in terms of branches)• Number of bank branches in 15 towns are as
follows:
1
38
2
7
3
30
4
32
5
17
6
28
7
26
8
19
9
26
10
26
11
37
12
24
13
33
14
29
15
28
Group 8 Assignment: Take a sample of 10 branches from a total of 500 branches spread
over 15 towns so that probability of selecting a town is proportional to size of the town
(measured in terms of branches)• Number of bank branches in 15 towns are as
follows:
1
38
2
7
3
30
4
32
5
17
6
28
7
26
8
19
9
26
10
26
11
37
12
24
13
3
14
59
15
28
Group 9 Assignment: Take a sample of 10 branches from a total of 500 branches spread
over 15 towns so that probability of selecting a town is proportional to size of the town
(measured in terms of branches)• Number of bank branches in 15 towns are as
follows:
1
3
2
7
3
30
4
32
5
17
6
28
7
26
8
19
9
26
10
26
11
37
12
24
13
38
14
59
15
28
Group 10 Assignment: Take a sample of 10 branches from a total of 500 branches spread
over 15 towns so that probability of selecting a town is proportional to size of the town
(measured in terms of branches)• Number of bank branches in 15 towns are as
follows:
1
3
2
57
3
30
4
32
5
17
6
28
7
26
8
19
9
26
10
26
11
37
12
24
13
38
14
9
15
28