RESEARCH DESIGN K. SYED ABUDAHEER, M.P.T., : RESEARCH DESIGN K. SYED ABUDAHEER, M.P.T., Slide 2: A research design is the arrangement of conditions for collection and analysis of data in a manner that aims to combine relevance to the research purpose with economy in procedure. (Claire Selltiz, 1962)
The research is the plan, structure, and strategy of investigations of answering the research question is the overall plan or blueprint the researchers select to carry our their study. Slide 3: Research design is always gives the answers of following questions
What is the study about ?
Why is the study being made ?
Where will the study be carried out ?
What type of data is required ?
Where can the required data be found ?
What periods of time will the study include ?
What will the sample design ?
How will the data be analyzed ? NEED FOR RESEARCH DESIGN : NEED FOR RESEARCH DESIGN Minimizes time and money
Selection of appropriate tools
Eliminate bias and marginal error FEATURES OF GOOD DESIGN : FEATURES OF GOOD DESIGN It should be flexible, appropriate, efficient, economical and so on.
It should give a smallest experimental error and high reliability and validity.
Good research design includes following five important elements.
Investigator’s role PRINCIPLES OF RESEARCH DESIGN : PRINCIPLES OF RESEARCH DESIGN Professor Fisher has enumerated three important principles of research design.
Principle of Replication
Principle of randomization
Principle of Local Control PRINCIPLE OF REPLICATION : PRINCIPLE OF REPLICATION According to this principle, the experiment should be repeated more than once. Thus, each treatment is applied in many experimental units instead of one.
By doing this method, the accuracy and precision of the study are increased significantly.
For example, the effect of two variety of rice. PRINCIPLE OF RANDOMIZATION : PRINCIPLE OF RANDOMIZATION This principle provides protection
This principle indicates that the researcher should design or plan the experiment in such a way that the variations caused by extraneous factors can all be combined under the general heading of “Chance”.
Example : effect of two variety of rice PRINCIPLE OF LOCAL CONTROL : PRINCIPLE OF LOCAL CONTROL The extraneous factors, the know source of variability, is made to vary deliberately over as wide a range as necessary and this needs to be done in such a way that the variability it causes can ne measured and hence eliminated from the experimental error.
Example : effect of two variety of rice TYPES OF RESEARCH DESIGN : TYPES OF RESEARCH DESIGN Experimental design - which obey the all three principles.
Quasi –Experimental design
Non- experimental design EXPERIMENTAL DESIGN : EXPERIMENTAL DESIGN The investigator planning an experiment has many experimental design option to choose. Experimental designs fall into two major categories.
True or Classical experimental design
Pre- experimental design TRUE OR CLASSICAL EXPERIMENTAL DESIGN : TRUE OR CLASSICAL EXPERIMENTAL DESIGN There are three major subdivisions in true or classical experimental design
Pre test – post test control group design
Solomon four group design
After / post test only experimental design PRE TEST – POST TEST CONTROL GROUP DESIGN : PRE TEST – POST TEST CONTROL GROUP DESIGN Experimental group
Pre test Post test
Pre test Post test PRE TEST – POST TEST CONTROL GROUP DESIGN : PRE TEST – POST TEST CONTROL GROUP DESIGN In this design, subjects have been designed randomly to the experimental or control group
The experimental treatment is given only to those in the experimental group, and the pre tests and post tests are those measurements of the dependent variables that are made before and after the experimental treatment is performed.
All true experimental designs have subjects randomly assigned groups, have an experimental treatment introduced to some of the subjects and have the effects of the treatment observed. advantages : The investigator is able to account for events occurring between time 1 and time 2 through observation of control group
It also enables the investigator to control for changes in the instrumentation, since changes or drifts in measurement should affect both groups equally
Randomization decreases selection bias and maturation. advantages SOLOMON FOUR GROUP DESIGN : SOLOMON FOUR GROUP DESIGN Experimental group I
Pre test Post test
Control group I
Pre test Post test
Experimental group II
Control group II
Post test SOLOMON FOUR GROUP DESIGN : SOLOMON FOUR GROUP DESIGN This design employs two experimental groups and two control groups. Initially, the investigator randomly assigns subjects to the four groups. Those in the first experimental treatment, and observed again on occasion 2.
Those in the experimental group 2 also receive the treatment but are observed only after the treatment, nor before.
Those in control group 1 are observed, on occasion 1 and 2, but they are not given the experimental treatment.
Those in control group 2 are observed only on the second occasion without previous observation or treatment. ADVANTAGES : ADVANTAGES It has great potential for generating information about differential sources of effect on the dependent variable, because all four groups are studied at the same time, both the effects of events occurring between time 1 and time 2 and the maturation of subjects are controlled.
One can examine the score of control groups 2 for a measure of maturation without the influence of treatment.
The invigilator also can compare the different in the groups. AFTER / POST TEST ONLY CONTROL GROUP DESIGN : AFTER / POST TEST ONLY CONTROL GROUP DESIGN Experimental group
Post test AFTER / POST TEST ONLY CONTROL GROUP DESIGN : AFTER / POST TEST ONLY CONTROL GROUP DESIGN This design, which is sometimes called after only control group design.
This is composed on two randomly assigned groups, but neither of which is pretested or premeasured in the before period of time.
The independent variable introduced into experimental group and withheld from the control group. ADVANTAGE : ADVANTAGE This design can be useful in situation where it is not possible to pretest the subjects or pretest is non essential PRE- EXPERIMENTAL DESIGN : PRE- EXPERIMENTAL DESIGN It is one type of the experimental design
It have three subdivisions. They are
One short case study or single case study
One group pretest – posttest design
The static group comparison design ONE SHORT CASE STUDY OR SINGLE CASE STUDY : ONE SHORT CASE STUDY OR SINGLE CASE STUDY Experimental treatment
In single case study, that studies at once, following a treatment or an agent presumed to cause change.
Because the study design has a total absence of control, it is considered to be little value as an experiment. ONE GROUP PRETEST – POSTTEST DESIGN : ONE GROUP PRETEST – POSTTEST DESIGN Experimental treatment
Pre test post test
Here only one group is observed before and after the independent variable is introduced.
Loss of the control group decreases the usefulness of the study but may be necessary in cases where it is not possible or feasible to have control groups. THE STATIC GROUP COMPARISON DESIGN : THE STATIC GROUP COMPARISON DESIGN The static group that has experienced the independent variable is compared with one that has not.
Here the experimental group received the independent variable, but control group did not receive the independent variable. QUASI EXPERIMENTAL DESIGN : QUASI EXPERIMENTAL DESIGN It is one which full experimental control, usually randomization is not possible.
It has three subdivisions. They are
Non equivalent control group design or the four celled design without use of randomization
The time series quasi experimental design
The multiple time series design Non equivalent control group design or the four celled design without use of randomization : Non equivalent control group design or the four celled design without use of randomization Experimental group ( not randomly selected)
Pre test Post test
Control group ( not randomly selected)
Pre test Post test time series experimental design : time series experimental design experimental treatment
Pre test 1 2 3 4 5 6 6 5 4 3 2 post test 1
The time series experiment design, a single group experiment comprises of series of observation in the before time period to establish a baseline. The experimental variable is then introduced, followed by another series of observation to examine the effect of the independent variable. The multiple time series design : The multiple time series design Experimental group
Pre test 1 2 3 4 5 6 6 5 4 3 2 post test 1
Pre test 1 2 3 4 5 6 6 5 4 3 2 post test 1 Non-experimental Research Designs : Non-experimental Research Designs This type of design is other wise called as weakest design.
Need not obey any principles
It have so many subdivisions Reasons for Undertaking Non experimental Studies : Reasons for Undertaking Non experimental Studies Number of human characteristics/ independent variables are not subject to experimental manipulation or randomization
Some variables cannot ethically be manipulated
For some research, it is not practical to conduct a true experiment/manipulate variables
For some situations, it is more realistic to explore phenomena in more natural manner
Nonexperimental research is often needed to scope out the experimental one Classification of Non Experimental designs : Classification of Non Experimental designs Descriptive/exploratory survey studies
Ex post facto studies
Cross-sectional & longitudinal studies
Retrospective & prospective studies Descriptive/Exploratory Survey : Descriptive/Exploratory Survey Broadest category
Detailed observations, descriptions & documentation of existing variables
Little is known about the phenomenon
Justifies, assesses current conditions/practice
Variables of interest: opinions, attitudes or facts
Determines differences between variables
Data collected by questionnaire or interview
Researchers only relate one variable to another, no attempt to determine causation Exploratory Survey Research : Exploratory Survey Research Designed to obtain information about
interrelations of variables within a population
Census vs. sample surveys
Flexibility and broadness
Superficiality – extensive vs. intensive analysis Two aspects about surveys : Two aspects about surveys Survey Content
Answering how, what,and to what extent questions
Usually focus on what was done and what people plan to do in the future Survey Administration
Different data collection methods
Self-administered questionnaires (SAQs)
Mixed-mode strategy Descriptive Research : Descriptive Research Purpose is to
observe, describe, & document aspects of a situation as it naturally occurs
serve as a starting point for hypothesis generation or theory development
Descriptive Correlational Studies
Univariate Descriptive Studies
Incidence Studies Descriptive Correlational Studies : Descriptive Correlational Studies Describes the relationship among variables rather than infer cause-and-effect relationships
Are usually cross-sectional Univariate Descriptive Studies : Univariate Descriptive Studies Could focus on one or more variables
Undertaken to describe the frequency of occurrence of a behavior or condition or each variable rather than relationships between or among them
Incidence Studies Prevalence Studies : Prevalence Studies Done to determine the prevalence rate of some condition at a specific point in time
Data is obtained from the population at risk for the condition – cross sectional design
Prevalence Rate (PR) = # cases with condition X K
# in population at risk Incidence Studies : Incidence Studies Used to measure the frequency of developing new cases
Need longitudinal designs
Incidence Rate (IR)=
# new cases with condition over given period X K
# at risk of becoming a new case
Relative Risk: an estimate of risk of “caseness” in one group vs. another; contribution of risk factors
E.g. males vs females for acquiring depression Descriptive/Exploratory Survey (cont’d) : Descriptive/Exploratory Survey (cont’d) Advantage:
large amount of information can be
obtained from a large population in an
economical manner which is “surprisingly” accurate
Info tends to be superficial as breadth is emphasized
Expertise in: sampling techniques, questionnaire construction, interviewing and data analysis to produce a reliable and valid study.
Time-consuming & sometimes costly Interrelationship/Difference Studies: Correlational Studies : Interrelationship/Difference Studies: Correlational Studies Examines if variables covary
Quantifies the strength or relationship between the variables (not cause & effect)
+ve or –ve direction relationship determined
Increased flexibility when investigating complex relationships among variables
Efficient and effective method of collecting a large amount of data
Potential for practical application in clinical settings
Potential foundation for future, experimental studies
Framework for exploring relationships that are not manipulated.
Unable to manipulate variable of interest
No randomization in sampling
Generalizability decreased as dealing with preexisting groups
Unable to determine a causal relationship because of the lack of manipulation, control and randomization. Interrelationship/Difference Studies: Ex Post Facto Studies : Interrelationship/Difference Studies: Ex Post Facto Studies Literally means ‘from after the fact’
Also known as causal-comparative studies or comparative studies
Explores differences/relationships between variables (similar to quasi-experimental designs)
Allows for establishment of a differential effect
Similar to correlational designs
Offers a higher level of control
Lack of control on variables
Unable to draw causal linkage
Problem of alternative hypothesis Interrelationship/Difference Studies: Developmental Studies : Interrelationship/Difference Studies: Developmental Studies Not only concerned with existing status & interrelationship of phenomena but also with changes from elapsed time.
Cross sectional (one/more time points, perhaps different groups) vs Longitudinal (several time points with same group over extended period)
Retrospective (dependent variable has already been affected by independent variable, link present events to past events) vs Prospective (link present events to presumed future effect, less common, considered stronger design) Others : Others Natural Experiments: study of a group exposed to natural or other phenomenon that have health or other consequences, compared with a nonexposed group; people are affected at random
Path Analytic Studies: using a technique called path analysis, nonexperimental data is tested against a hypothesized causal inference Epidemiologic Research Designs : Epidemiologic Research Designs Cohort studies: trend study in which specific subpopulations (eg age specific) are examined over time for generational differences
prospective & retrospective
Case-control study: comparison of cases/subjects (with specific condition), with controls (without condition); only difference should be exposure to presumed cause
Cross-sectional design: phenomena under study are captured during one period of data collection; one point in time