Research Design Blue print for conducting a study APRIL 21, 2014 RG 701- ADVANCE RESEARCH METHODS
What is Research Design? A plan, structure and strategy of investigation so conceived as to obtain answers to research question or problem.
Blueprint of research study Operationalizing the variables Sample selection Data collection Analyzing results
‘How’ of the research journey Once Research question has been finalized ◦Research Problem: decided what you want to study THEN Need to determine how to conduct study Need to design your research ◦Set of instructions for researcher to gather and analyze data ◦Scheme of plan of actions for meeting the objectives
Purpose To provide a plan in answering research question Important Consideration: ◦To minimize the possible errors ◦To maximize the validity and reliability of data In order to maximize the control over factors that could interfere with the validity of the findings
Functions of a Research Design 1.Conceptualize an operational plan to undertake the various procedures and tasks required to complete your study 2.Control of variance: Ensure that these procedures are adequate to obtain valid, objective and accurate answers to the research question
The researcher’s plan Highlights the methods and tools that are used during research process Name ‘study design’ Detailed information regarding: ◦How the study will be conducted ◦Study population ◦Sample size ◦Type of data that will be collected ◦Identification of variables ◦Data collection means/method ◦Where interviews (if required) be conducted ◦Managing Ethical issues ◦Etc.
Reliability of Variables Consistency, stability, or dependability of data A research method should yield the same result, even if conducted twice or more
Validity of variables Data need not only be reliable but valid as well Validity refers to which extent an instrument is able to actually measure what it is supposed to measure
2. Theory of Causality Selection of a study design that will minimize the effects of ◦extraneous variables influencing the independent variable ◦chance or random error: due to random or chance variables –change in dependent variable due to respondent’s state of mode or ambiguity in the research instrument (ambiguous questions)
STUDY DESIGN IN QUANTITATIVE RESEARCH
Types of Research Design See Fig 8.1of text book Classification based on: ◦Number of contacts ◦Reference period of study ◦Nature of the investigation
Number of Contacts Three or moreTwoOne X-sectional studies Before and after studies Longitudinal studies
Reference Period Retrospective - Prospective Prospective Retrospective
Nature of Investigation Semi-experimentalNon-experimental Experimental
Study Based on NUMBER OF CONTACTS
Cross-sectional Study Design Simple in design Also known as ‘one-shot’ or ‘status studies’ Best suited for finding out the prevalence of a phenomenon, situation, problem, attitude or issue Takes one time cross-section of a population Cross-sectional with regard to both ◦Study population ◦Time of investigation Disadvantage: ◦Can not measure change
Before & After Studies Also known as ‘Pre-test/post-test’ design Measures change Can be described as two sets of cross-sectional data collection points on the same population in time Change is measured by comparing the difference in the variable(s) before and after intervention Disadvantages ◦Expensive and difficult ◦Measures total change including due to extraneous var. ◦Time lapse problems ◦Reactive effect of instrument ◦Regression effect
Longitudinal Studies Before and after only measures change but not the trend… Longitudinal studies determine the pattern in change in relation to time Intervals may vary from study to study Considered as series of repetitive cross-sectional studies Allows researcher to measure the pattern of change with enhanced accuracy Disadvantages ◦Conditioning effect
Study Based on REFERENCE PERIOD TIME FRAME IN WHICH STUDY IS EXPLORING A PROBLEM
Retrospective Investigates a phenomenon, situation, problem or issue that has happened in the past conducted on the basis of: ◦data available for that period or ◦respondents’ recall of the situation
Prospective Refers to the likely prevalence of a phenomenon, situation, problem or outcome in the future Establishes the outcome of an event or what is likely to happen Researcher must wait for an intervention to register its effects
Retrospective -Prospective Focuses on past trends in a phenomenon and study it into the future Part of data is collected retrospectively from existing records before the intervention in introduce Influence of intervention is studied afterwards
Study Based on NATURE OF INVESTIGATION
Experimental Starting from cause to establish its effects Introducing the intervention and monitor the change Flexibility of controlling or manipulating the independent variable Some commonly used experimental study designs: ◦Before-and-after ◦Control group design ◦Double control design ◦Comparative design
Control group design Two population groups are selected for comparison (comparable in every respect except for the intervention) : 1.Experimental group ◦Receives/is exposed to the intervention 2.Control group ◦not exposed to the intervention 3.Chief objective of the control group is to quantify the impact of extraneous variables 4.Helps to ascertain the impact of the intervention only
Continu….Control group design 1.Make ‘Before’ observations on both groups (same time) 2.Experimental group is exposed to intervention 3.Make ‘after’ observations on both groups 4.Calculate difference in the ‘before’ and ‘after’ observations between the groups regarding dependent variable(s)
Study population Intervention I/var. Experimental Group Control Group D/var. Y’e D/var. Y”e D/var. Y’c D/var. Y”c Exp G: Total Change in D/var. Ye = (Y”e-Y’e) = impact of (intervention ± extraneous var. ± chance var.) Cont. G: Total Change in D/var. Yc = (Y”c-Y’c) = impact of (extraneous var. ± chance var.) Difference between Ex & Cont Groups equals to the ‘Impact of the Intervention’
Double Control Groups Two control groups To quantify the impact due to the research instrument Exclude one control group from ‘before’ observation
Comparative design The study of the effectiveness of different interventions The study population is divided into same number of groups as the number of interventions
Matched Control Experimental Design Identical members from study population selected for study
Non-experimental Starting from effects to trace the cause effects/outcomes causation Observation of a phenomenon and then establishing its cause
Semi-experimental Has properties of both experimental and non-experimental studies
References Text book