RESEARCH DESIGNS
Sanchita Garai, Mathangi Sumana Sri and Sanjit Maiti
RESEARCH DESIGN
Concept
• Research design is the detailed plan of an
investigation
• It is the blue print of the detailed procedure of
testing the hypotheses and analysis of the obtained
data
Importance of Research Design
• How to collect data for testing the hypotheses
• Arrive at a valid, objective, accurate & economic
solution of the given problem
• Control variables which may influence the research
outcome
• Which variables should be treated as control
variables, what methods of manipulation will be
relevant in a particular situation
• What types of statistical analyses should be
performed
TYPES OF RESEARCH DESIGNS
Exploratory or formulative research design
The main purpose of this design is to formulate a problem for more
precise investigation or to develop working hypothesis. Pilot study is
conducted prior to the main investigation.
Techniques used in conducting Exploratory Research Study
• Secondary data analysis
• Experience survey or Pilot survey
• Focus group interview
TYPES OF RESEARCH DESIGNS
Descriptive and Diagnostic research design
A descriptive research design is used for fact
finding with adequate interpretation. Diagnostic
research design is concerned with the testing of
hypothesis and interpreting relationship between
the variables.
This design serves a variety of research
objectives:
• Description of phenomena or characteristics
associated with the population
• Discovery of associations among different
variables and in case of causal relationship studies
• Discovery and measurement of cause-and-effect
relationships among variables
Steps to be followed in Descriptive and Diagnostic
Research Design
1. Selection of the problem
2. Formulation of the objectives of the study
3. Formulation of hypothesis
4. Selection and development of tools for data collection
5. Defining the population and selection of sample
6. Preparation of structured interview schedule/ questionnaire
7. Pretesting the interview schedule/ questionnaire
8. Determination of the reliability and validity of the instruments developed
9. Collection of data
10. Processing and analysis of data
11. Findings and discussions
12. Preparation of report
TYPES OF RESEARCH DESIGNS
Hypothesis-testing or Experimental research design
It is the blue print of procedures that enables the researcher to test
hypothesis by reaching valid conclusions about relationships between
independent and dependent variables. In descriptive research study, no
control as well as manipulation of variables are required but in
experimental study some degree of control over variables and
manipulation of independent variables, to study their effect on
dependent variables are desirable.
Principles of experimental research design (Prof R.A. Fisher)
• Principle of Replication
• Principle of Randomization
• Principle of Local control
Principles of Experimental Research Design
• Principle of Replication: Deliberate repetition of an experiment, using
nearly identical procedure, which may sometimes be with a different set
of subjects in a different setting and at different time periods.
• Principle of Randomization: Technique in which each member of the
population or universe has an equal and independent chance of being
selected.
• Principle of Local control: Refers to the amount of balancing, blocking
and grouping of subjects or the experimental units employed in the
research design.
Grouping- assignment of homogenous subjects or experimental units into
a group
Blocking- assignment of experimental units to different blocks in such a
way that the assigned experimental units within a block may be
homogenous
Balancing- grouping, blocking and assignment of experimental units to the
different treatments in such a way that the resulting design appears to be a
balanced one
Types of Experimental Research Designs
True experimental designs:
1. Posttest only, Equivalent-group design
2. Pretest-Posttest control group design
Quasi-experimental designs:
3. Time-series design
4. Equivalent time samples design
5. Non-equivalent control group design
Ex-post facto designs:
6. Ex-post-facto cause to effect design
7. Ex-post-facto effect to cause design
TRUE EXPERIMENTAL RESEARCH DESIGNS
• Researchers have complete control over extraneous variables
• Can predict confidently that the observed effect on the dependent
variable is only due to the manipulation of the independent variable
Essential characteristics of true experimental research:
1. Manipulation
2. Control
3. Randomization
• Two equivalent groups of the same size are randomly selected from a
population
• One group, randomly given the treatment, is experimental group and the
other group for which no treatment is given, is called the control group
• After the treatment, posttests are administered to both the groups and the
difference is measured
• Also known as After only design
Posttest only, Equivalent Group Design
Example: Impact of the farm advisories during COVID-19 pandemic on
the mental health of the farming community in the Haryana state of India
Posttest only, Equivalent Group Design
• This design is similar to the posttest only equivalent group design,
except for the fact that it also makes a provision for pretest of both
experimental and control groups before the treatment is administered
• Also known as Before-after design
• This design provides for following measurements and comparisons:
1. Change in score of experimental group due to treatment
2. Change in score of control group due to natural causes
3. Effect of treatment, by taking into consideration (controlling) both the
initial condition (pretest score) and the impact due to natural causes
Pretest-posttest Control Group Design
Example: Effectiveness of a mobile app on the adoption of the eco-
friendly dairy farming by the farmers in Haryana state of India
Pretest-posttest Control Group Design
QUASI-EXPERIMENTAL RESEARCH DESIGNS
• Quasi-experimental research design involves the
manipulation of independent variable to observe the
effect on dependent variable, but it lacks at least one of
the two characteristics of the true experimental design,
randomization or a control group
• In other words, quasi-experimental designs have an
element of manipulation but lack at least one of the other
two properties that characterize true experiments,
randomization or control group
• These designs are generally used to establish the
causality (effect of independent variable on dependent
variables) in situations where researchers are not able to
randomly assign the subjects to groups or for various
reasons no control group is available for an experimental
study
• A series of pretests are given or pre treatment measurements are made of
the selected group or equivalent groups
• Then the treatment is administered and a series of posttests are given or
post treatment measurements are made of the selected group or
equivalent groups
• A control group or comparison group is not included in this design
• This design is more applicable where testing is a regular feature of the
setting, such as educational institutions, or where data are regularly
collected, such as records of production, cost of living indices, etc
Time-series Design
Example: Comparison of performance of students in the final examination
for a number of years before and after the introduction of a new method of
teaching or evaluation
SELECTION OF A GROUP
SERIES OF PRETESTS OR PRETREATMENT MEASUREMENTS
OF THE SAME GROUP OR EQUIVALENT GROUPS AT
REGULAR INTERVALS
TREATEMENT
SERIES OF POSTTESTS OR POSTTREATMENT
MEASUREMENTS OF THE SAME GROUP OR EQUIVALENT
GROUPS AT REGULAR INTERVALS
Time-series Design
• This design is an improvement over and extension of time-series
design, because of repeated introduction of treatments, followed by
posttests every time in a systematic way
• A single group is used and there is no control group
• Overall change and change separately with different treatments can
be measured
• As there is carry over experience from one treatment to the other, it
cannot be specifically stated which treatment produced what effect
Equivalent Time-samples Design
Example: Effect of NICRA Project on the attitude change of farmers
towards climate resilient practices in Haryana state of India
SELECTION OF A GROUP
TREATMENT-A, POSTTEST-A
TREATMENT-B, POSTTEST-B
TREATMENT-C, POSTTEST-C
TREATMENT-D, POSTTEST-D
Equivalent Time-samples Design
• The design is similar to pretest posttest control group design, except that
the method of randomization cannot be applied in the assignment of
subjects to experimental and control groups
• To find out treatment effect, the difference between the posttest and
pretest scores of the experimental and control groups may be computed
Non-Equivalent Control Group Design
Example: Impact of climate resilient agricultural practices in
improving livelihood of the farmers of Coastal Andhra Pradesh
Non-Equivalent Control Group Design
EX-POST-FACTO RESEARCH DESIGNS
• Researches in which the events have taken place prior to undertaking
the investigation fall under this category of research design
• Investigator has no scope to manipulate the independent variables as
they have already occurred
• Inferences on the relationships between independent and dependent
variables are drawn on the basis of effects already manifest
• In this design, the treatment is determined not by manipulation but by
selection
• In this design, a worthwhile cause is first identified and then its effects
are traced out
• The presumed cause is regarded as the independent variable
(treatment) and the presumed effect as the dependent variable
• Matching of groups is an important aspect of this design
• Unless the experimental and control groups are properly matched on
important variables, the investigator may come out with erroneous
conclusions
• The random selection of respondents makes the design valid
Ex-post-facto Cause to Effect Design
Example: Impact of TV viewing on the change in Knowledge and
Adoption of farmers in the Haryana state of India
Ex-post-facto Cause to Effect Design
• This is a reverse process of cause to effect design, in which the
presumed cause is traced from the presumed effects
• As in the earlier design, matching of groups is important
• Random selection of respondents makes the design valid
Ex-post-facto Effect to Cause Design
Example: Extent to which levels of knowledge, motivation and
adoption of farmers could be attributed to TV viewing
Ex-post-facto Effect to Cause Design
Research Design  for  social science.pptx

Research Design for social science.pptx

  • 1.
    RESEARCH DESIGNS Sanchita Garai,Mathangi Sumana Sri and Sanjit Maiti
  • 2.
    RESEARCH DESIGN Concept • Researchdesign is the detailed plan of an investigation • It is the blue print of the detailed procedure of testing the hypotheses and analysis of the obtained data Importance of Research Design • How to collect data for testing the hypotheses • Arrive at a valid, objective, accurate & economic solution of the given problem • Control variables which may influence the research outcome • Which variables should be treated as control variables, what methods of manipulation will be relevant in a particular situation • What types of statistical analyses should be performed
  • 3.
    TYPES OF RESEARCHDESIGNS Exploratory or formulative research design The main purpose of this design is to formulate a problem for more precise investigation or to develop working hypothesis. Pilot study is conducted prior to the main investigation. Techniques used in conducting Exploratory Research Study • Secondary data analysis • Experience survey or Pilot survey • Focus group interview
  • 4.
    TYPES OF RESEARCHDESIGNS Descriptive and Diagnostic research design A descriptive research design is used for fact finding with adequate interpretation. Diagnostic research design is concerned with the testing of hypothesis and interpreting relationship between the variables. This design serves a variety of research objectives: • Description of phenomena or characteristics associated with the population • Discovery of associations among different variables and in case of causal relationship studies • Discovery and measurement of cause-and-effect relationships among variables
  • 5.
    Steps to befollowed in Descriptive and Diagnostic Research Design 1. Selection of the problem 2. Formulation of the objectives of the study 3. Formulation of hypothesis 4. Selection and development of tools for data collection 5. Defining the population and selection of sample 6. Preparation of structured interview schedule/ questionnaire 7. Pretesting the interview schedule/ questionnaire 8. Determination of the reliability and validity of the instruments developed 9. Collection of data 10. Processing and analysis of data 11. Findings and discussions 12. Preparation of report
  • 6.
    TYPES OF RESEARCHDESIGNS Hypothesis-testing or Experimental research design It is the blue print of procedures that enables the researcher to test hypothesis by reaching valid conclusions about relationships between independent and dependent variables. In descriptive research study, no control as well as manipulation of variables are required but in experimental study some degree of control over variables and manipulation of independent variables, to study their effect on dependent variables are desirable. Principles of experimental research design (Prof R.A. Fisher) • Principle of Replication • Principle of Randomization • Principle of Local control
  • 7.
    Principles of ExperimentalResearch Design • Principle of Replication: Deliberate repetition of an experiment, using nearly identical procedure, which may sometimes be with a different set of subjects in a different setting and at different time periods. • Principle of Randomization: Technique in which each member of the population or universe has an equal and independent chance of being selected. • Principle of Local control: Refers to the amount of balancing, blocking and grouping of subjects or the experimental units employed in the research design. Grouping- assignment of homogenous subjects or experimental units into a group Blocking- assignment of experimental units to different blocks in such a way that the assigned experimental units within a block may be homogenous Balancing- grouping, blocking and assignment of experimental units to the different treatments in such a way that the resulting design appears to be a balanced one
  • 8.
    Types of ExperimentalResearch Designs True experimental designs: 1. Posttest only, Equivalent-group design 2. Pretest-Posttest control group design Quasi-experimental designs: 3. Time-series design 4. Equivalent time samples design 5. Non-equivalent control group design Ex-post facto designs: 6. Ex-post-facto cause to effect design 7. Ex-post-facto effect to cause design
  • 9.
    TRUE EXPERIMENTAL RESEARCHDESIGNS • Researchers have complete control over extraneous variables • Can predict confidently that the observed effect on the dependent variable is only due to the manipulation of the independent variable Essential characteristics of true experimental research: 1. Manipulation 2. Control 3. Randomization
  • 10.
    • Two equivalentgroups of the same size are randomly selected from a population • One group, randomly given the treatment, is experimental group and the other group for which no treatment is given, is called the control group • After the treatment, posttests are administered to both the groups and the difference is measured • Also known as After only design Posttest only, Equivalent Group Design Example: Impact of the farm advisories during COVID-19 pandemic on the mental health of the farming community in the Haryana state of India
  • 11.
  • 12.
    • This designis similar to the posttest only equivalent group design, except for the fact that it also makes a provision for pretest of both experimental and control groups before the treatment is administered • Also known as Before-after design • This design provides for following measurements and comparisons: 1. Change in score of experimental group due to treatment 2. Change in score of control group due to natural causes 3. Effect of treatment, by taking into consideration (controlling) both the initial condition (pretest score) and the impact due to natural causes Pretest-posttest Control Group Design Example: Effectiveness of a mobile app on the adoption of the eco- friendly dairy farming by the farmers in Haryana state of India
  • 13.
  • 14.
    QUASI-EXPERIMENTAL RESEARCH DESIGNS •Quasi-experimental research design involves the manipulation of independent variable to observe the effect on dependent variable, but it lacks at least one of the two characteristics of the true experimental design, randomization or a control group • In other words, quasi-experimental designs have an element of manipulation but lack at least one of the other two properties that characterize true experiments, randomization or control group • These designs are generally used to establish the causality (effect of independent variable on dependent variables) in situations where researchers are not able to randomly assign the subjects to groups or for various reasons no control group is available for an experimental study
  • 15.
    • A seriesof pretests are given or pre treatment measurements are made of the selected group or equivalent groups • Then the treatment is administered and a series of posttests are given or post treatment measurements are made of the selected group or equivalent groups • A control group or comparison group is not included in this design • This design is more applicable where testing is a regular feature of the setting, such as educational institutions, or where data are regularly collected, such as records of production, cost of living indices, etc Time-series Design Example: Comparison of performance of students in the final examination for a number of years before and after the introduction of a new method of teaching or evaluation
  • 16.
    SELECTION OF AGROUP SERIES OF PRETESTS OR PRETREATMENT MEASUREMENTS OF THE SAME GROUP OR EQUIVALENT GROUPS AT REGULAR INTERVALS TREATEMENT SERIES OF POSTTESTS OR POSTTREATMENT MEASUREMENTS OF THE SAME GROUP OR EQUIVALENT GROUPS AT REGULAR INTERVALS Time-series Design
  • 17.
    • This designis an improvement over and extension of time-series design, because of repeated introduction of treatments, followed by posttests every time in a systematic way • A single group is used and there is no control group • Overall change and change separately with different treatments can be measured • As there is carry over experience from one treatment to the other, it cannot be specifically stated which treatment produced what effect Equivalent Time-samples Design Example: Effect of NICRA Project on the attitude change of farmers towards climate resilient practices in Haryana state of India
  • 18.
    SELECTION OF AGROUP TREATMENT-A, POSTTEST-A TREATMENT-B, POSTTEST-B TREATMENT-C, POSTTEST-C TREATMENT-D, POSTTEST-D Equivalent Time-samples Design
  • 19.
    • The designis similar to pretest posttest control group design, except that the method of randomization cannot be applied in the assignment of subjects to experimental and control groups • To find out treatment effect, the difference between the posttest and pretest scores of the experimental and control groups may be computed Non-Equivalent Control Group Design Example: Impact of climate resilient agricultural practices in improving livelihood of the farmers of Coastal Andhra Pradesh
  • 20.
  • 21.
    EX-POST-FACTO RESEARCH DESIGNS •Researches in which the events have taken place prior to undertaking the investigation fall under this category of research design • Investigator has no scope to manipulate the independent variables as they have already occurred • Inferences on the relationships between independent and dependent variables are drawn on the basis of effects already manifest • In this design, the treatment is determined not by manipulation but by selection
  • 22.
    • In thisdesign, a worthwhile cause is first identified and then its effects are traced out • The presumed cause is regarded as the independent variable (treatment) and the presumed effect as the dependent variable • Matching of groups is an important aspect of this design • Unless the experimental and control groups are properly matched on important variables, the investigator may come out with erroneous conclusions • The random selection of respondents makes the design valid Ex-post-facto Cause to Effect Design Example: Impact of TV viewing on the change in Knowledge and Adoption of farmers in the Haryana state of India
  • 23.
  • 24.
    • This isa reverse process of cause to effect design, in which the presumed cause is traced from the presumed effects • As in the earlier design, matching of groups is important • Random selection of respondents makes the design valid Ex-post-facto Effect to Cause Design Example: Extent to which levels of knowledge, motivation and adoption of farmers could be attributed to TV viewing
  • 25.