Research Design:
the overall structure or plan of the research project
your plan to answer the research question.
Research Methodology:
The study of the scientific steps adopted by a researcher in conducting a research.
Largely theoretical.
Research Methods:
The steps and techniques used for conduction of research, especially data collection and analysis.
It has a narrower scope than methodology.
2. DISCLAIMER!
• Research is very practical; the more you do the more you know.
• This lecture is largely introductory; the fine details are beyond the scope of this lecture.
• Oversimplification for ease of appreciation; there are always nuances.
• Tailored for medical students
• Engaging a more experienced researcher (a.k.a. advisor or supervisor) is instrumental
to a successful research.
• There are a number of chat box exercises in this lecture. Correctly and swiftly
answering one of them comes with a gift.
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
3. KEY DEFINITIONS
• Research Design:
– the overall structure or plan of the research project
– your plan to answer the research question.
• Research Methodology:
– The study of the scientific steps adopted by a researcher in conducting a research.
– Largely theoretical.
• Research Methods:
– The steps and techniques used for conduction of research, especially data collection and
analysis.
– It has a narrower scope than methodology.
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
4. THE SCIENTIFIC METHOD
• The defining feature of scientific research and hence its design.
– Empirical approach
– Observations
– Questions (answerable)
– Hypotheses (null and alternate/experimental)
– Experiments (or research study)
– Analyses
– Conclusion
– Replication
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
6. CORRELATION VS. CAUSATION
• CORRELATION
– Simply a relationship between two variables.
– May be Positive: variables go in same direction or Negative: variables go in opposite
directions
– May allow for prediction e.g. predicting what levels of high BP is associated with stroke
from positive correlational studies
– Does NOT equal causation.
• CAUSATION
– Includes correlation, PLUS
– time-order relationship and
– NO alternative explanations.
• Ice-cream sales is correlated with homicides in a city.
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
7. Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
8. QUALITATIVE VS. QUANTITATIVE
RESEARCH
• QUANTITATIVE RESEARCH
– Quantitative research makes use of statistical analyses to obtain findings.
– Formal, systematic measurement plus statistics.
– A game of NUMBERS
– E.g. cross-sectional study
• QUALITATIVE
– No mathematics; just WORDS
– Involves interviews or observations without formal measurement
– Often serves as a source of hypotheses for further quantitative research.
– E.g. case study
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
9. OBSERVATIONAL VS. EXPERIMENTAL
STUDIES
• OBSERVATIONAL
– Just looks.
– Observes responses
– No influence on, or interferences with the responses
– Does NOT determine causation, but can establish a relationship.
• EXPERIMENTAL
– Observes responses after interfering with variables
– Can determine causation
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
10. STATISTICAL BIAS
• When a statistic is not representative of the population
– Selection
– Recall
– Observer
– Social desirability
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
11. CHAT BOX EXERCISE:
WHAT DO YOU THINK OF THE
ACCURACY AND PRECISION IN
DARTBOARD B?
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
12. THE GOALS OF RESEARCH
• EXPLORATION/FORMULATION
– Discovery of ideas or insights; little or no prior knowledge
– Mostly qualitative – little need for precision/accuracy
– Requires flexibility in design - for consideration of the different facets to a problem.
• DESCRIPTION/DIAGNOSIS
– Describes characteristics and association/correlation
– Mostly quantitative – precision/accuracy are very important
• EXPLANATION/CAUSATION
– Tests hypotheses on causal relationship between variables
– Permits inferences on causality
– Least bias and highest reliability
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
13. WHAT THEN IS RESEARCH DESIGN?
• Research design is the framework or blueprint for a particular research project.
• It is a plan to answer the research question.
• And so it depends on and follows the definition of the research problem.
• It clarifies what? Where? When? How much? By what means? as far as the research is
concerned.
• Includes an elaborate outline/description of the steps that will be taken from
hypothesizing to data analysis and the resultant discussion/conclusion
• It includes the methods of the research
• Enables easy progress of the research.
• Most results are gotten out of the project with the least possible resources.
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
14. WHAT MAKES FOR A GOOD RESEARCH
DESIGN?
• Not a one size fits all.
• CONSIDER:
– Purpose of study. E.g. A descriptive study would require a design high accuracy; study for
causal relationship would require the least bias and most reliability; explorative studies
ought to be very flexible, etc.
– Resources
– Skills / know-how
• In general: minimizes error and bias, maximizes reliability and information
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
15. DETAILS OF A RESEARCH DESIGN
• Where are we conducting this study? Study setting
• What is the study design?
• What population are we conducting this study on? Study population
• What is the number of people we are studying? Sample size
• How do we obtain our sample? Sampling Technique
• How will the data be collected from this sample? Data collection
• How should we analyze the obtained data? Data Analysis
• How should we report our results? Data presentation
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
16. PRELIMINARIES
• CHOOSING A RESEARCH TOPIC
• LITERATURE REVIEW
• FORMULATING A RESEARCH PROBLEM
– Use operational definitions
• ESTABLISHING HYPOTHESES
– Null & Alternate
– Falsifiable & Predictive
• DETERMINING THE VARIABLES
– Independent vs. Dependent Variables
– Categorical vs. Continuous Variables
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
17. CHAT BOX EXERCISE:
WHAT TYPE OF STUDY:
DESCRIPTIVE, EXPLORATORY OR
EXPLANATORY?
You want to investigate a group of factory
workers who are suspected to have
environmentally developed lupus. What
type of study would you use?
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
18. STUDY SETTING
• Relevant geography of the area
• Relevant socio-demographic and cultural characteristics
• Occupation
• Overall health-status
• Healthcare system
• Population
• Details given here must be RELEVANT to the proposed study.
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
19. POPULATION
• Target or theoretical population
• Study population or accessible population
– Subset of target population
– Source of samples
– Eligibility criteria
– Sampling method
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
20. Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
21. Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
22. Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
23. SAMPLE SIZE DETERMINATION
• Choosing the number of observations that will be in the sample for study.
• Based on resources, convenience and the need for enough statistical power.
• Ingredients:
– The effect size: the size of difference () to detect minimal effect of interest.
– Standard deviation SD or () of the data – the variability.
– The desired power () of the experiment to detect the postulated effect (80% or 90%)
– The significance level ( =0.05 or 0.01)
– The Z score (from confidence level) –E.g. 95% CL – z score of 1.96
• Larger sample sizes are generally encouraged: increased precision, increased power.
• Formulae exists especially for normally distributed statistics.
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
25. DESCRIPTIVE (OBSERVATIONAL)
STUDIES
• Cross-sectional
– Prevalence study
– “snapshot” – data is gotten from a population at a specific point in time.
– NO information on causality (no temporality) or result of intervention.
– Cost effective, but prone to (recall) bias.
– May be descriptive or analytical
• Case report / case study
• Case series
• Ecologic studies
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
26. ANALYTICAL (OBSERVATIONAL)
STUDIES
• Case control
– Odds ratio
• Cohort
– Prospective vs. retrospective
– Relative risk
• Cross – sectional studies
– Aims to establish association
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
27. Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
28. OTHERS…
• Systematic review
– Gathering, appraising and synthesizing evidence to answer a well defined question.
– Presented as meta-analysis, or
– Textual summary
• Meta analysis
– Statistical procedure from combining numerical data from different studies.
– Follows systematic review
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
30. INTERVENTIONAL STUDIES CONTD…
• Experimental studies
– Randomized control /clinical trial
• Randomization
• Blinding
• Placebo
• Quasi-experimental (non-randomized) studies
– Field trials
– Community trials
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
31. CHAT BOX EXERCISE:
WHAT TYPE OF STUDY DESIGN IS
THIS RESEARCHER CONDUCTING?
A researcher wants to look at the relationship between
tobacco smoking in adolescence and lung cancer in
adulthood. Every adult living in Calabar is contacted and
asked if they smoked tobacco as a child and whether they
have lung cancer now.
What type of study design is this researcher conducting?
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
32. DATA COLLECTION
• Questionnaires
– Self vs. interviewer administered
– Open vs. closed ended questions
• Focus Group Discussion
– Group perception, attitudes, beliefs, opinions etc.
– Participants, moderator, note taker.
• In-Depth Interviews
– Deeper insight on individual’s perspective.
• Data sheet and Clinical records – case notes, registers etc.
• Observational checklists
– Facilities, treatment practices etc.
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
33. DATA COLLECTION…
• Training of interviewers
• Informed consent / assent
• Pretesting
– Using questionnaire in a small pilot study to determine how well it works.
• Pilot study
– Smaller version of a larger study to ascertain feasibility of protocol, and fish out
weaknesses.
• Confidentiality
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
34. MEASUREMENT SCALE
• Scales of measurement are ways to sub-categorize different types of data.
• Qualitative/categorical
– Nominal: with or without order; dichotomous
– Ordinal
• Quantitative (continuous)
– Interval
– Ratio
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
35. Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
36. DATA MANAGEMENT
• Data screening
– Obtained data is checked for errors, and errors are fixed
• Data entry
– Labelling and coding
– Computerized input into the system
• Data analysis
– Using relevant software applications. E.g. SPSS.
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
37. DATA ANALYSIS
• Makes meaning of the obtained data
• Depends largely on statistics
• Descriptive statistics: describe the characteristics and establish relationships
• Inferential statistics: determine causal relationships and draw conclusions about the
population from which the sample was drawn
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
38. DESCRIPTIVE STATISTICS
• Frequency Distribution
• Measures of central tendency
– Mean
– Median
– Mode
• Measures of dispersion
– Standard deviation
– Variance
• Measures of correlation
– Depends on the type of scale: nominal, ordinal, interval or ratio
– Pearson, spearman rank, etc.
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
39. INFERENTIAL STATISTICS
• T test
– Independent
– paired
• ANOVA
• Chi-square
• Regression
– Simple or multiple
– Linear or Logistic regression
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
40. DATA PRESENTATION
• Tables
• Charts
– Pie charts
– Bar charts
– Histograms
• Scatter plots
• Graphs
• Pictorial
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
41. Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
42. CHAT BOX EXERCISE:
WHAT TYPE OF STUDY DESIGN IS
THIS RESEARCHER
CONDUCTING?
A researcher is studying predictive factors for failed induction of
labour. He enrolls patients into two groups: A = patients who had
failed inductions, and B = patients who had successful induction.
Physiologic data measured prior to induction of labour are
collected from the case notes and analyzed to see which data
differed between the two groups.
What type of study design is this researcher conducting?
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
43. TAKE HOME
• Carefully choose the research topic
• Use operational definitions
• Articulate hypotheses that are falsifiable or predictive
• Use the most appropriate research design
• Use random sampling or assignments whenever possible
• Fight against bias
• Measure with reliable and valid techniques
• Analyze and interpret data with care
• Remain ethical.
• Find the real source of your happiness and pursue it… it doesn’t have to be research though.
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
44. FURTHER READING…
• C.R. Kothari. Research Methodology: Methods and techniques. New Age Publications
(Academic)
• Peter Pruzan. Research Methodology: The Aims Practices and Ethics of Science
• Jason L.A, Glenwick D.S. Handboook of methodological approaches to community-
based research: qualitative, quantitative and mixed methods. Oxford University Press.
• Essentials of Research Design and Methodology by Geoffrey R. Marczyk, David
DeMatteo, David Festinger.
Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)
45. Research Design and Methodology - Mazpa Ejikem
(NiMSA SCORE - TORASIF Webinar)