Professional Documents
Culture Documents
Prem Prasad Panta: Research and Biostatistics
Prem Prasad Panta: Research and Biostatistics
(Associate Professor)
Research and Biostatistics
Karnali Academy of Health Sciences
Objectives
Commonly used active verbs
Quntitative research Qualitative research
• Descriptive research • Ethnography
– Identify, Assess, Find out – Assess, describe, explain,
– Describe, Compare discover
• Correlational • Grounded theory
– Determine, examine, – Develop, extend,
identify
identify, validate
• Experimental
• Phenomenological
– Determine , examine,
investigate, measure, – Describe, develop,
detect express
Hypothesis
• Tentative statement of the expected relationship between two or
more variables
Types of hypothesis
• Simple and complex
– Simple hypothesis: relationship between 2 variables
– Complex hypothesis: relationship between more than 2 variables
• Associative and causal hypothesis
– Associative hypothesis: association between variables at natural settings
– Causal hypothesis: predicts the cause and effect relationship
• Directional and non directional
– Directional: direction is specified, positive and negative relationship
– Non directional: not specify the direction, there is relationship
• Null and alternative hypothesis
– Null hypothesis: statistical hypothesis: no relationship
– Alternative hypothesis: research hypothesis: there is relationship
Literature Review
• Defined as broad, comprehensive, in depth,
systematic, and critical review of scholarly
publication, unpublished printed, or audio
visual materials and personal
communications.
• It develops new knowledge, insights and
overall scenario of that topic
Types of literature review
On the basis of purposes and methodologies, it would be broadly
classified as:
• Traditional and narrative literature review: gather a volume of
literature ad summarize the facts and findings,
• Systematic literature review : more rigorous and more
systematic, check reliability and validity of research, selection
criteria, quality of research, inclusion and exclusion criteria for
literature, analysis part
• Meta-analysis: advanced form of systematic literature review,
involves the findings of several quantitative studies on a single
subject area and draw the integrated inferences and conclusions
• Meta-synthesis: non statistical technique, used in qualitative
research, integrates the research of findings of grounded theory,
phenomenological and ethnographic studies
Sources of literature
• Published information
- Primary source: Original, peer-reviewed and published
research journal articles by original researchers
- Secondary sources: description of studies prepared by
someone other than the original researcher, comment
and summaries of multiple research studies; e.g. Meta-
analysis, systematic reviews and meta- synthesis
• Unpublished information
• Experts, scientific societies and organisations
6
Resource of literature
Journals
Magazines
and news Books
papers
Resources of
literature
Research review Electronic
report database
Conference
Thesis
papers
Electronic Sources
• Search engine: Google, Yahoo, MSN etc.
• Using database:
– Pubmed (http://www.ncbi.nlm.nih.gov/pubmed)
– ResearchGate (www.researchgate.net
– Web of Science (https://www.webofknowledge.com)
– Cochrane (systematic review for health sciences related
literature)
– WHO countries (free database, WHOSIS=who statistical
information system)
– African Index Medicus (African biomedical information, support
by WHO)
• Google Scholar
• Using online version of the Journals
• Online magazines, e-books, etc.
8
Research Design
• Experimental research design
– True experimental design/RCT
– Quasi experimental design
– Pre experimental design
• Features
– Intervention/Manipulation
– Randomization
– control
Observational research design
• Descriptive research design: accurate description of
characteristics of individual/group, natural setting, no
manipulation, no control, variables are studied as they
exit in the real world.
– Case report
– Case series
– Cross sectional
• Case control : causes of diseases is investigated
• Cohort: longitudinal approach, investigation of
occurrence of disease in a cohort with existing
presumed causes
Research design
• Meta analysis: Quantitatively combining and integrating
the findings of multiple research studies on a particular
topic
• Secondary data analysis: collected data are reanalyzed by
others
• Evaluation studies: involves the judgement about success
of programs, practices, procedures or policies
• Operational research: involves the study of complex
human organizations and services to develop new
knowledge about institutions, programs, use of facilities
and personnel in order to improve working efficiency of an
organization.
Randomization
• Means that every subject has an equal chance
of being assigned to the experimental or
control group.
• Also know as random assignment of subjects
• Used in true experimental research design
• Minimize the threat of internal validity of the
study and to eliminate the effect of
extraneous variable on dependent variable
Descriptive research
• Case report: Descriptive observational study
of a single patient or case with rare, specific
and unique features
• Case series: group of patients with similar
clinical presentation, diagnosis or who are
exposed to similar intervention are observed
over a period of time
Cross sectional research
• Relationship between disease and other
factors at one point in time in defined
population
• Prevalence studies or epidemiological studies
• Exposure and outcome measured at the same
point of time
• Types: univariant, explorative and comparitive
Case control research
• Causes of a disease are investigated after the
occurrence of a disease
•Association between exposure and outcome= odds ratio
•Specially used for studying rare and emerging diseases
•Not possible to calculate the incidence or prevalence rate
Cause Effect
Cohort research design
• Cohort: group of people who have something in common
and who remain part of a group over an extended time
• Longitudinal approach is used to investigate the
occurrence of disease in a cohort
• Known as incidence studies
• Individuals subjects are followed over time to measure
the exposure the exposure when it happens, then they
measure the outcome at a point in time after exposure
• Used to measure the risk differences, risk ratio,
somewhere odds ratio
• Also known as prospective study
Cohort research
Cause Effect
True experimental design
• Effect of independent variable is observed on dependent variable
• Used to find the cause and effect relationship
• Independent variable is manipulated through treatment or
intervention
• Randomization and control
Example
• Post test only control design
• Pretest-post test only design
• Factorial design
• Randomized block design
• Latin square design
Quasi- Experimental Research
• Popular method in health science research
where experimental design is not done
• Cause and effect is studies
• Manipulation is done in independent variable
but at least two properties are absent
(randomization or control group )
Sample size
• To measure the prevalence
n = Z2pq/d2
Z = 1.96 at 95% CI
P = prevalence of an attribute
Q=1-P
D= allowable error (5% to 10%)
• To estimate the average
n= Z2SD2/d2
SD = standard deviation
Data collection
Data sources:
• Primary sources
– People, objects, programs etc.
– Collected through interviews, questioning,
observations, measurements
• Secondary sources:
– Published: journals, newspapers, government
reports, census reports, mass communications
– Unpublished: thesis, dissertations, official records
Types of method
Techniques Tools
Interview: Interview schedule/ Questionnaire
Questioning(self report): Questionnaire/attitude scale
Observation: Rating scale/check list/video
Types of Question
Open format(open ended)
Closed format(closed ended): Dichotomous/Multiple
choice/cafeteria/rank order/rating/importance/ Likert
question
Reliability
• Reliability means the consistency with which set of
test scores measure whatever they do measure.
• The reliability of the measuring scale must be stable,
consistent.
• Reliability is necessary for validity but not sufficient.
• A measurement is reliable it gives the same results
in many times
• Reliability refers to the repeatability of findings.
• A instrument is reliable if it has absence of error.
Validity
• The most critical criteria of the scaling
technique is validity.
• A scale is said to be valid when it measures
correctly what is expected to be measure.
Reliability and Validity
Reliability Validity
• It is the extent to which • A valid measure is one
the measurements which is measuring what
resulting from a test it is supposed to measure.
reflect characteristics of • A valid measure must be
the subject of reliable but a reliable
measurement measure may not be valid.
• An experiment is reliable • Validity refers that getting
is we are getting results reflect the real
consistent result from one.
the same measure
Reliability and Validity
• Reliability: How consistent or stable is the
instrument?
– Is the instrument dependable?
• It is broadly classified as
a. Predictive validity : it is the degree of
forecasting judgment
b. Concurrent validity: It is the degree of the
measures at present
4. Construct Validity
• Construct validity is "the degree to which a test
measures what it claims, or rationales, to be
measuring.
• Construct validity is the extent to which the
measurements used, often questionnaires,
actually test the hypothesis or theory they are
measuring. Construct validity should
demonstrate that scores on a particular test do
predict the theoretical trait it says it does.
Interpretation
• The validity coefficient is measured in terms of
the correlation coefficient (r) of the scores of
he different tests. If r lies in between
• 0.9 ≤ r ≤ 1 =Very high validity
• 0.8 ≤ r ≤ 0.9 =high validity
• 0.6 ≤ r ≤ 0.8 =Satisfactory validity
• 0.4 ≤ r ≤ 0.6 =moderate validity
• 0.0 ≤ r ≤ 0.4 =poor validity
Format of a research report
• Principle of ethics
– Beneficence: Positive risk benefit, focus on
potential risk and physical harm; avoid discomfort
– Justice: fair and nondiscriminatory selection,
anonymity and confidentiality, not misused of
data, protect the vulnerable group
– Respect of human dignity: right to question and
clarification, right to quit from study, fully
informed consent
Confidentiality and Anonymity
• Confidentiality and anonymity are ethical practices
designed to protect the privacy of human subjects
while collecting, analyzing, and reporting data.
• Confidentiality refers to a condition in which the
researcher knows the identity of a research subject,
but takes steps to protect that identity from being
discovered by others.
• Anonymity is a condition in which the identity of
individual subjects is not known to researchers.
Respect for the Dignity of Persons
• Respect for the autonomy of persons –
Promoting the essential freedom in decision making
attributable to persons based on their moral and
rational dignity and their capacity for self-
determination
• The active protection of persons –
Requiring those who are dependent or vulnerable
be affordable security against harm or abuse.
• No research should prevail over for human rights,
fundamental freedoms and human dignity.
• Right to full disclosure
• Inform consent
Beneficence (Non-maleficence)
• Obligation to maximize possible benefits and to
minimize possible harms and wrongs (Do Good).
13. To analyze the relationship between blood pressure and height in a sample of
students, we would use
• Paired t-test
• Correlation coefficient
• Chi square test
• Independent t-test
14. Which of the following is a non-parametric test?
a. ANOVA test
b. Student t test
c. Chi square test
d. Z test
15. While computing chi-square test for a 3x2 contingency table, then degree of
freedom is………………….
e. One
f. Two
g. Three
h. Four
MCQ
16. Which of the following is not a major parts to the
research report?
a. Abstract
b. Results
c. Method
d. Footnote
17. Which of the following is non-probability sampling?
a. Snowball
b. Random
c. Cluster
d. Stratified
Multiple Choice Question
18. A cross sectional study design is also known as
a. Incidence study
b. Prevalence study
c. Relative risk study
d. None of the above
19. Internal validity refers to ..
e. Whether or not there is really a causal relationship between two
variables
f. Whether or not the findings are relevant to the participants’
everyday lives
g. The degree to which the researcher feels that this was a worthwhile
projects
h. How accurately the measurements represents underlying concepts
20. The key defining characteristic of experimental
research is that:
a. The independent variable is manipulated
b. Hypotheses are proved
c. A positive correlation exists
d. Samples are large
21. In an experiment, the group that does not
receive the intervention is called:
a. The experimental group
b. The participant group
c. The control group
d. The treatment group
22. Standard deviation is a measure of:
a. range
b. central tendency
c. inference
d. dispersion
23. In an experiment the researcher manipulates the
______variable and measure the ___ variable.
a. independent, dependent
b. dependent, independent
c. causal, spurious
d. spurious, causal
24. In a normal distribution most scores are located:
a. at the high end of the distribution
b. at the low end of the distribution
c. in the center of the distribution
d. at the extremes of the distribution