RESEARCH METHODOLOGY

&
BIOSTATISTICS

* Few jewels from ocean

Dr. Kusum Gaur
Professor, PSM
WHO Fellow IEC
 Definition of Research

“Research is a
systematized effort
to gain new knowledge”.

12/08/2012 Dr. Kusum Gaur 2

 Steps in Research (Holy 11)

1. Collect review of literature/Situation Analysis

2. Identify and prioritize health problems
3. Decide aims & objectives
4. Planning Methodology
5. Execution
6. Compilation, Classification & Presentation of data
7. Analysis
8. Test of Significance/Test of Hypothesis
9. Inferences
10.Report Writing
11. Dissemination of Report

12/08/2012 Dr. Kusum Gaur 3

 Process of Concluding
8 7 6

Reporting Inferences Analysis

Data Collection
5

Steps in Research

Execution

Execution
Research Problem
Define
1

for Pretest
Collection
Data
Review of Literature Methodology
4

2 3
Planning

12/08/2012 Dr. Kusum Gaur 4

 STEP-1

DEFINITION
OF THE
RESEARCH PROBLEM

12/08/2012 Dr. Kusum Gaur 5

 RESEARCH PROBLEM ?

Research Problem refers to some difficulty

which a researcher experiences and
wants to obtain a solution for the same.

i.e. a question or issue to be examined.

12/08/2012 Dr. Kusum Gaur 6

 Process of Defining Problem

Analysis of the Situation

Identify & Prioritize Problems

Select & Define Problem

Statement of
Research Objectives

12/08/2012 Dr. Kusum Gaur 7

 CRITERIA OF SELECTION
The selection of one appropriate researchable
problem out of the identified problems requires
evaluation of certain criteria.

* Internal / Personal criteria – Researcher’s side

* External Criteria – Problem side factors

12/08/2012 Dr. Kusum Gaur 8

INTERNAL CRITERIA OF SELECTION

 Researcher’s Interest,

 Researcher’s Competence,

 Researcher’s own Resource:

 Human Resource
 Money
 Material
 Time

12/08/2012 Dr. Kusum Gaur 9

EXTERNAL CRITERIA OF SELECTION

 Researchability of the problem,

 Importance and Urgency,
 Novelty of the Problem,
 Feasibility,
 Facilities,
 Social Relevance
 Public health Importance

12/08/2012 Dr. Kusum Gaur 10

 DEFINE RESEARCH PROBLEM
(Title of the Research Topic)

 Transforming the selected research problem into a

scientifically researchable statement.

 Problem definition or Problem statement should be

clear, precise, self-explanatory and include:-

 What
 How
 When
 Where

12/08/2012 Dr. Kusum Gaur 11

 RESEARCH OBJECTIVES
(Objectives)
 Research Objectives are the statement of the
questions that is to be investigated with the goal of
answering the overall research problem.

 Research Objectives should be clear and achievable.

 Generally, they are written as statements, using the

word “to”
(For example, ‘to discover …’, ‘to determine …’, ‘to
establish …’, ‘to find out -----’, ‘to assess -----’etc. )

Objectives should infer in the end of the study

12/08/2012 Dr. Kusum Gaur 12

 Hypothetical Research Question
 Problem:
PCR of Diabetes Mellitus is increasing very fast
during last five year

 Mission:
Reduce the incidence of heart disease

 Belief:
Meditation is good to reduce stress which is an
important precursor of DM

 Hypothesis
H- Meditation decreases the risk of DM
12/08/2012 Dr. Kusum Gaur 13
 Association of Garlic consumption with
coronary Artery Diseases

Aim: To Study the association of Meditation with

Diabetes Mellitus in patients attending at Medical OPD
of SMS Hospital, Jaipur (Raj) India.

Objectives:
1. To assess and compare the proportion of DM cases in
individuals doing regular meditation and not doing
meditation.

2. To find out the risk ratio of DM in individuals not

doing meditation on doing regular meditation.
 STEP-2

REVIEW
OF
LITERATURE

12/08/2012 Dr. Kusum Gaur 15

 Review of literature

 What ?

 Why ?

 Where ?

12/08/2012 Dr. Kusum Gaur 16

 What ?
REVIEW OF LITERATURE

Literature Review is the documentation

of a published and unpublished work

from secondary sources of data

in the areas of specific interest to the researcher.

12/08/2012 Dr. Kusum Gaur 17

 Why ? - PURPOSE OF REVIEW
 Tofind out already investigated problems and
those that need further investigation.

 To formulate researchable hypothesis.

 To gain a background knowledge

 To identify data sources

 To learn how others structured their reports.

12/08/2012 Dr. Kusum Gaur 18
 Where ?
SOURCES OF LITERATURE
 Books and Journals
 Databases
Bibliographic Databases
Abstract Databases
Full-Text Databases
 Govt. and NGO Records & Reports
 Internet
 On line journals: ww.articalbase.com …….
 E. Databases – Popline, Medline …….
 Research Dissertations / Thesis

12/08/2012 Dr. Kusum Gaur 19

 Step-3

Methodology
 Methodology
Study Area : Location of study - Hospital, community etc.

Study Period: Start to end of Study (maximum period available for

study should be defined)

*Selection of Study Design

* Selection of Study Population

Pre-requisits of study: Study Tools, Terminologies, Orientation

trainings etc.

*will be taken separately

12/08/2012 Dr. Kusum Gaur 21

 Methodology……
• Study Tools for data collection: subjects, proforma,
examination, measurements, lab investigations
• Planning
 Data collection, compilation, data entry
 Data cleaning
 Analysis plan:
• Confidentiality
• Ethical clearance: Consent from
 Institutional Review Board
 Observational units

12/08/2012 Dr. Kusum Gaur 22

 Study Design

A study design is a specific plan or protocol

for conducting the study,
which allows the investigator
to translate the conceptual hypothesis
into an operational one.

12/08/2012 Dr. Kusum Gaur 23

 Direction of Study

Backward Forward

Cross -sectional

Retrospective Prospective

4. Ambidirectional
12/08/2012 Dr. Kusum Gaur 24
 Decision Tree
Intervention Done
No Yes
Observational Study Experimental Study

Comparison Group Randomization

No Yes
No Yes
Descriptive Study Analytic Study
NRCT Study RCT Study

Direction of Study

E O E O
Cohort Study E = O Case-Control Study
Cross-Sectional Study

12/08/2012 Dr. Kusum Gaur 25

 Epidemiological Study Design
Observational Studies
 Descriptive Studies

 Analytic
Cross-Sectional
Case-Control
Cohort

Experimental / Interventional studies

 As per Control: RCT/NRCT
 As per Blinding: Single /Double Blind  
 As per Design: Simple/Cross-over
 As per Area: Field/Clinical/Lab
12/08/2012 Dr. Kusum Gaur 26
 Descriptive Studies

• Case reports
• Case series
• Population studies

12/08/2012 Dr. Kusum Gaur 27

 Descriptive Studies: Uses

• Hypothesis generating

• Suggesting associations

12/08/2012 Dr. Kusum Gaur 28

 Descriptive Type of Observational Study

• Other Name Case-Series/Population

• Unit of Study Case/Individuals
• Study Question What is happening 
• Direction Of Inquiry
• Study Design
☻☻☻☻☻☻ desired information
☻☻☻☻☻☻ about cases/individuals is
collected

12/08/2012 Dr. Kusum Gaur 29

 Case-Series …….

Advantages
• Easy to do
• Excellent at identifying unusual situation
• Good for generating hypotheses

Disadvantages
• Generally short-term
• Investigators self-select (bias!)
• no controls
09/03/2010 Dr. Kusum Gaur 30
Analytical Observational Studies

• Cross-sectional

• Case-control

• Cohort

12/08/2012 Dr. Kusum Gaur 31

 Cross-sectional Study
• Data collected at a single point in time

• Describes associations

• Prevalence
A “Snapshot”

12/08/2012 Dr. Kusum Gaur 32

 Cross-Sectional Study

• Other Name Prevalence Study

• Unit of Study Individual
• Study Question What is happening 
• Direction of Inquiry
•  Study Design Exposed
to Factor

Not
 Exposed
Diseased to Factor

Population Exposed to
 Factor
Non-
Disease Not
Exposed to
12/08/2012 Dr. Kusum Gaur Factor 33
 Objectives of a Cross-Sectional
Study

To find out association

12/08/2012 Dr. Kusum Gaur 34

 Cross-sectional Study

Sample
DefinedofPopulation
Population

Regular Not doing meditation

Meditation

Prevalence of Prevalence of
DM DM

Time Frame = Present

12/08/2012 Dr. Kusum Gaur 35
 Cross-sectional Study
E.G. Out of 1000 population if 100 were doing meditation regularly &
out of that only 2 were having DM. Remaining 900 were not doing
meditation at all, out of that 220 were having DM.

+ DM -

+ 2 98
Meditation

- 220 680

12/08/2012 Dr. Kusum Gaur 36

 Cross-Sectional Study

• Strengths
– Quick
– Cheap

• Weaknesses
– Cannot establish cause-effect

09/03/2010 Dr. Kusum Gaur 37

 Case-Control Studies
 Start with people who have disease(Cases)

 Match them with controls that do not have

disease (Match Confounding)

 Look back and assess exposures

12/08/2012 Dr. Kusum Gaur 38

 Controls

A control is a standard of comparison

(confounded with variability but without effect)

for

• Effects

• Variability

12/08/2012 Dr. Kusum Gaur 39

 Case-Control Study
• Other Name Retrospective Study
• Unit of Study Cases/Control
• Study Question What has happened 
• Direction of Inquiry= F O
•  Study Design
Exposed

 Cases
Not
Exposed

Exposed

Control
Not
Exposed
12/08/2012 Dr. Kusum Gaur 40
Objective of a Case-Control Study

To find out association

To assess Risk Ratio

12/08/2012 Dr. Kusum Gaur 41

 Case-Control Study

Cases
Regular Meditation
Patients with DM
No Meditation

Controls
Regular Meditation
Persons w/o DM
No Meditation

Past Present
12/08/2012 Dr. Kusum Gaur 42
The logic of Case-Control Studies

 Cases differ from controls only in having the

disease

 If exposure does not predispose to having the

disease, then exposure should be equally
distributed between the cases and controls.

 The extent of greater previous exposure among

the cases reflects the increased risk that
exposure confers
12/08/2012 Dr. Kusum Gaur 43
Case-Control Studies: Strengths

• Good for rare outcomes: cancer

• Can examine relation of exposures to disease
• Useful to generate hypothesis
• Fast
• Cheap
• Provides Odds Ratio

09/03/2010 Dr. Kusum Gaur 44

Case-Control Studies: Weaknesses

• Cannot measure
– Incidence
– Prevalence
– Relative Risk

• Can only study one outcome

• High susceptibility to bias

09/03/2010 Dr. Kusum Gaur 45

 Cohort Study

• Begin with disease-free individuals

• Classify patients as exposed/unexposed

• Record outcomes in both groups

• Compare outcomes using relative risk

12/08/2012 Dr. Kusum Gaur 46

 Cohort Study
• Other Name Prospective Study / Follow-up Study/Incidence Study
• Unit of Study Individual
• Study Question What is happening 
• Direction of Inquiry F O
•  Study Design Diseased
•
Exposed to Not Non
Factor Diseased

Cohort
Cohort Diseased
Not
Exposed to
Factor
Non-Diseased

12/08/2012 Dr. Kusum Gaur 47

 Logic of Cohort Study

 Cohort is a group of persons sharing a

common characteristics

 Differences in the rate at which exposed and

control subjects contract a disease is due to
the differences in exposure, since others are
known and similar.

12/08/2012 Dr. Kusum Gaur 48

 Cohort Study

 Prospective (usually)

 Controlled

 Can determine causes and incidence of

diseases as well as identify risk factors

 Generally expensive, time consuming and

difficult to carry out
12/08/2012 Dr. Kusum Gaur 49
 Steps for Cohort Study

 Identify geographically defined group

 Identify exposed subjects and not exposed
subjects
 Follow over a specific time
 Record the fraction in each group who
developthe condition of interest
 Compare these fractions using RR, AR or OR

12/08/2012 Dr. Kusum Gaur 50

 Objectives of a Cohort Study

 To find out association

 To assess Risk Ratio

 To find out Relative Risk

 To find out Attributed Risk

12/08/2012 Dr. Kusum Gaur 51

 Prospective Cohort Study
DM
No Meditation
No DM

Cohort
DM
Regular
Meditation No DM

Present Future
12/08/2012 Dr. Kusum Gaur 52
 Cohort Study: Strengths

• Can measure multiple outcomes

• Can adjust for confounding variables

• Can calculate Attributed Risk

09/03/2010 Dr. Kusum Gaur 53

 Cohort Study: Weaknesses

• Expensive

• Time consuming
• Cannot study rare outcomes
• Confounding variables

09/03/2010 Dr. Kusum Gaur 54

 Measurements of association

Cohort Study Case Control Study

•Significance Test •Significance Test

•Relative Risk •OR
•Attributable Risk
•OR

12/08/2012 Dr. Kusum Gaur 55

 Measures of Association

Significance Test – to test significance of

difference in exposure between control and
Cases
Odds ratio - ratio of the odds of contracting
disease in given exposure
Relative Risk – Ratio between incidence
among exposed and incidence among non-
exposed
Attributed Risk – percentage of difference
between incidence among exposed and non-
exposed with incidence among exposed
RR or OR of 1 indicate no effect of exposure (equal odds)
12/08/2012 Dr. Kusum Gaur 56
 ‘Z’ Score of Exposure Rates

Cases control

Exposed a b
a x 100
Exposure Rates = in Cases (P2) Non- c d
exposed
a+c

b x 100
Exposure Rates = in Controls (P1) P2 – P1

b+d Z Score =
SEDP

P1Q 1 P2Q2
09/03/2010 Dr. Kusum Gaur
SEDP = ------------- + -------- 57
N1 N2
 ad
ODD’s Ratio = Times
bc

Incidence among Exposed

RR = Times
Incidence among Non-Exposed

a/a+b a (c+d)
= =
c/c+d c (a+b)

09/03/2010 Dr. Kusum Gaur 58

 Attributed Risk

(Incidence among Exposed - Incidence among Non-Exposed)

AR = x 100
Incidence among Exposed
a
Incidence among Exposed= x 100
a+b
c
Incidence among Non-Exposed= x 100
c+d
09/03/2010 Dr. Kusum Gaur 59
 Experimental Studies

Clinical trials provide the “gold standard” of

determining the relationship between factor

and the event

12/08/2012 Dr. Kusum Gaur 60

Types of Experimental Study

As per Randomization:
• Randomized Control Trials (RCT)

• Concurrent Parallel Design (RCT)

• Sequential RCT Design

• RCT with External Control

• Non – Randomized Trials (NRCT)

12/08/2012 Dr. Kusum Gaur 61
Types of Experimental Study….

As per Design:
• Simple

• Cross-Over Study Design

As per Study Area:

• Field Trials

• Clinical Trials

12/08/2012 • Lab. Trials Dr. Kusum Gaur 62

 Quality of Experimental Study

• Randomization

• Blinding

• Control

• Cross-Over

12/08/2012 Dr. Kusum Gaur 63

 Controls in Clinical Trials

A clinical trial is a comparative, prospective

experiment conducted in human subjects

• Historical controls are better than no

controls

• Patients can serve as own controls - This is

usually beneficial as the comparison
removes patient differences

12/08/2012 Dr. Kusum Gaur 64

 Blinding

Good practice: factors that can affect the

evaluation of outcome should not be permitted
to influence the evaluation process

Single-blind
Patient or evaluator (either of one) is blinded as
to intervention

Double-blind design
Neither patient nor outcome evaluator knows Rx
to which patient was assigned

12/08/2012 Dr. Kusum Gaur 65

 Randomized Control Trials (RCT)

• Before and After Comparison

• Comparison with Placebo

• Comparison Of two medicine/procedure/tests

• Comparison Of > two medicine/procedure/tests

12/08/2012 Dr. Kusum Gaur 66

 Experimental Study
• Other Name Intervention Study
• Objective To know the effect of intervention
• Unit of Study Individual meeting entry criteria
• Study Question What is happening after intervention in
both groups 
• Direction of Inquiry I E
• Study Design 1(Intervention with Placebo)
Positive
Outcome
 
Group 1/cases Intervention
Negative
Outcome

Positive
Outcome
Group
Placebo
2/control
Negative
Outcome

12/08/2012 Dr. Kusum Gaur 67

 Clinical Trial

R Treatment
a Outcomes
Group
n
d
Study o
Population m

i Outcomes
z Control Group
e

12/08/2012 Dr. Kusum Gaur 68

 Intervention Study - Design 2
(Comparison of Effect of Two Interventions)

Cases
Meeting
Entry criteria

Group - 1 Group -2

Intervention -1 Intervention Intervention - 2

Positive Negative Positive

Outcome Negative
Outcome Outcome Outcome

12/08/2012 Dr. Kusum Gaur 69

 Cross Over Design
Group -1 Cases Group-2
Meeting
Entry
criteria Intervention - 2
Intervention - 1

Positive Negative
Positive Negative Outcome
Outcome Outcome
Outcome

Group -1
Group -2 Crossover

Intervention -2
Intervention -1

Positive Negative
Positive Negative
Outcome
Outcome Outcome
Outcome

12/08/2012 Dr. Kusum Gaur 70

Other Types of Experimental Study

• Quincy Experimental Study

• Block Experimental Study

12/08/2012 Dr. Kusum Gaur 71

 Quincy Experimental Study

Cases Meeting
Entry criteria

Group - 1 Group -2

Intervention Intervention No Intervention

Positive Negative Positive

Outcome Negative
Outcome Outcome Outcome

12/08/2012 Dr. Kusum Gaur 72

 Block Experimental Study

Cases Meeting
Entry criteria

Group -3
Group - 1

Group -2

Intervention Intervention-3
Intervention -1 Intervention

Intervention-2

Positive Positive
Negative Negative
Outcome Outcome Outcome Outcome

Positive Negative
Outcome Outcome

12/08/2012 Dr. Kusum Gaur 73

Steps of Experimental Study
Drawing up a Protocol

Reference Population

Sample Population

Exclusions

Randomization
Experimental Group Control Group

Manipulation/Intervention

Follow - up

12/08/2012 Assessment
Dr. Kusum Gaur of Outcome 74
 Strength
Ideal Study Design for established causality

Weakness
Ethical Issues
STUDY QUESTIONS AND APPROPRIATE DESIGNS

Type of Question Appropriate Study Design

Burden of illness Field Surveys
- Prevalence Cross Sectional Survey
- Incidence Longitudinal survey

Causation, Risk & Prognosis Case Control Study,

Cohort study, RCT

Treatment Efficacy Randomized Controlled study

Diagnostic Test Evaluation Randomized Controlled study

Cost Effectiveness Randomized Controlled study

12/08/2012 Dr. Kusum Gaur 76

 Hierarchy of Epidemiological Study Design

Establish Causality RCT

Cohort

Case Control

Cross-Sectional

Case Series

Generate Hypothesis Case Report

12/08/2012 Dr. Kusum Gaur 77

 Methodology

Study Area : Location of study - Hospital, community etc.

Study Period: Start to end of Study (maximum period available for

study should be defined)

*Selection of Study Design

* Selection of Study Population

Sample Size
Sampling Technique

Pre-requisits of study: Study Tools, Terminologies, Orientation

trainings etc.

12/08/2012 Dr. Kusum Gaur 78

 Selection of study population

Whole Population

Sample Population

12/08/2012 Dr. Kusum Gaur 79

 What is Sample ?

• A sample is a small representative

segment of a population

• Inferences drawn from a sample are

expected to be applicable for the source
population

12/08/2012 Dr. Kusum Gaur 80

 Why do we need a sample?

To get inferences

applicable to universe

with minimum resources

12/08/2012 Dr. Kusum Gaur 81

 Sample – Qualities

Sample is a part of population but it is true

representative of whole.

Qualities

Adequate size

Appropriate sampling technique

12/08/2012 Dr. Kusum Gaur 82

 Factors on which SAMPLE SIZE depend:
• Population Factors
– Type of information available
• Type of study
– Type of Data
– Type of study design
– Type of sampling
– Type of Statistical Analysis for outcome needed
• Determined values of research by researcher
– Power
– Significance level

12/08/2012 Dr. Kusum Gaur 83

 Power: Ability to detect right answer

Alpha Error: Chance to miss right answer

 Type of Data & level of Measurements

Qualitative – Counted Facts – Nominal Data

Measured as Numbers expressed as proportions
 
Quantitative- Measured Facts - Numerical Data
Measured as quantity & expressed as Mean ± SD

*Ordinal Data – Rank Order Data

Measured as rank & expressed as Median ± Percentile

12/08/2012 Dr. Kusum Gaur 91

 Sample size for Qualitative data
 
Z 2 PQ 4 PQ
Sample Size= ------------------- -- = ------------------
L2 L2

P= Prevalence of disease
Q = 100-P
L = allowable error
Z= 1.96 ≈ 2 for 95% CL
for descriptive/case-series type of study design

09/03/2010 Dr. Kusum Gaur 92

 Sample size for Quantitative data
 

Z 2 SD 2 4 SD 2
Sample Size= ------------------- -- =----------------------
L2 L2

SD= Standard Deviation

L = allowable error
Z= 1.96 ≈ 2 for 95% CL
For Descriptive Studies only

09/03/2010 Dr. Kusum Gaur 94

 Finite Correction

Sample Size – Finite Population (where the

population is less than 50,000)
SS
New SS = _________________
( 1 + ( SS – 1 ))Pop
 How many controls?

n
k Here n0=No. of cases &
2n0  n n = expected no. of cases

• k = 13 / (2*11 – 13) = 13 / 9 = 1.44

• kn0 = 1.44*11 ≈ 16 controls (and 11 cases)
– Same precision as 13 controls and 13 cases
 Sampling Design factors of sample size

Variance of Specified Sampling

Design Effect =
Variance of Simple Random Sampling

12/08/2012 Dr. Kusum Gaur 97

Sampling Technique effect on Sample Size

Sampling Technique Design Effect Size Multiplier

Simple Random Sampling 1

Systemic Random Sampling 1.2

Stratified Random Sampling 0.8

Cluster Random Sampling 2

12/08/2012 Dr. Kusum Gaur 98

 Conventionally accepted
Researcher’s Estimations

Alpha Error 0.05

Power 80%

Confidence Limit 95%

12/08/2012 Dr. Kusum Gaur 99

 Key Concepts: Sample size
• Sampling Design - larger sample for Custer

• Desired Power – more power for larger sample

• Allowable error – smaller error for larger sample

• Heterogeneity leads to have larger sample to

cover diversities

• Nature of Analysis – Complex multivariate needs

larger sample
12/08/2012 Dr. Kusum Gaur 100
 Steps -Sample Size Estimation
• Stage 1- * Base Sample Size Calculation (n)

• Stage 2 – Sample Size with Design Effect (d)

=n*d

• Stage 3- Contingency Addition (e.g. 5%)

SS Estimation for study population =(n*d)
+5%of n

*Use appropriate equation for sample size

calculation
http://stat.ubc.ca/~rollin/stats/ssize/
12/08/2012 Dr. Kusum Gaur 101
E.G. Mean 1= 5, Mean 2 = 15 & SD = 14 inputting values
12/08/2012 Dr. Kusum Gaur 107
12/08/2012 Dr. Kusum Gaur 108
12/08/2012 Dr. Kusum Gaur 109
12/08/2012 Dr. Kusum Gaur 110
12/08/2012 Dr. Kusum Gaur 111
12/08/2012 Dr. Kusum Gaur 112
 SAMPLING
TECHNIQUES
 SAMPLING TECHNIQUES

• PROBABILITY/RANDOM SAMPLING

• NONPROBABILITY SAMPLING

12/08/2012 Dr. Kusum Gaur 119

 Random sampling Techniques

Aim is to give equal chance to

every observation unit to be
selected for study in sample.

(Any Observation unit

should not have Zero Probability )

12/08/2012 Dr. Kusum Gaur 120

 * Random Sampling Techniques

Simple Random Technique

Systemic Random Technique

Stratified Random Technique

Multiphase Random Technique

Multistage Random Technique

Cluster Random Technique

12/08/2012 Dr. Kusum Gaur 121

 Simple Random Technique

• Lottery Method

• Random Table Method

12/08/2012 Dr. Kusum Gaur 122

12/08/2012 Dr. Kusum Gaur 123
 Steps –Use of Random Table
• Stage 1- Give number to each member of population
• Stage 2 – Determine total population size (N)
• Stage 3- Determine Sample size (S)

• Stage 4 – Drop one finger on Random Table with eyes closed

• Stage 5 – Drop one finger with eyes closed on direction to be
chosen – Up/Down/Rt/Lt

• Stage 6- Determine first number within 0 to N

• Stage 7- * Determine other numbers till Sample size (S)

* Once a number is chosen do not repeat it again

12/08/2012 Dr. Kusum Gaur 124

 Steps –Use of Random Table..
e.g. N=300, M=50

Random no. Selected no. (3 digits from 0-300)

49468
49699
14043 043
15013 013
12600
33122 122
94169 169
89916
74169 169
32007 007
www.evaluation wikiog/index/how_to_use_a_random_number_Table
12/08/2012 Dr. Kusum Gaur 125
 Systemic Random Technique

The selection of sample follows a systematic

interval of selection
• Find serial interval
(K) = total population/sample size
• 1st observation through simple random sampling
th
among 1to K.
• Next observation = (1st +K) thObservation
• Next observation = (2nd +K) Observation
• -------------so on till No. of observations =
Sample Size

12/08/2012 Dr. Kusum Gaur 126

 Systemic Random Technique Population
N=100 (Given) 1 21 41 61 81
2 22 42 62 82
S=20 (Estimated) 3 23 43 63 83
K=N/S =100/20 =5 4 24 44 64 84
5 25 45 65 85
1st observation between 1 to 5 6 26 46 66 86
7 27 47 67 87
though SRS e.g. 3 8 28 48 68 88
Every 5th observation from 3rd 9 29 49 69 89
10 30 50 70 90
observation will be included in 11 31 51 71 91
sample population 12 32 52 72 92
13 33 53 73 93
So, sample population will be – 3rd 14 34 54 74 94
8th 13th 18th 23rd 28th 33rd 38th 43rd 48th 15 35 55 75 95
16 36 56 76 96
53rd 58th 63rd 68th 73rd 78th 83rd 88th 93rd 17 37 57 77 97
and 98th observation 18 38 58 78 98
19 39 59 79 99
20 40 60 80 100
12/08/2012 Dr. Kusum Gaur 127
 Stratified Random Technique
Sample selection through Simple Random/Systemic Random Technique

Sample Strata 1

Sample Strata 2

Strata 3
Sample

12/08/2012 Dr. Kusum Gaur 128

 Multiphase Random Technique
Specific test
Screening Test
S/S
Population

Probable cases Cases

Suspected cases For
study

12/08/2012 Dr. Kusum Gaur 129

 Multistage Random Technique

Each stage Simple RT is used village

district
village

village
State 1 district
Population village
Study
Of Population
Nation village
district
village
State 2
village
district
village

12/08/2012 Dr. Kusum Gaur 130

 Cluster Random Technique
The unit of random selection is a cluster rather than individual
• CI = Total population /30 (in 30 Cluster Technique)

Cluster 1 Cluster 27

Cluster 2 Cluster 28
Population Study
Of Population
Nation Cluster 3 Cluster 29

Cluster 30
Cluster 4

Through Simple RT

12/08/2012 Dr. Kusum Gaur 131

 Stratified Vs Cluster Technique

Stratified Technique Cluster Technique

• Homogenous groups • Comparable groups of
are made population are made
• Randomly select (usually 30)
sample from each group • Randomly select
• To make it more truly sample from each
representative, take group
sample population • More chances of error
proportion to size (PPS) than simple random
• Less chances of error
than simple random
 Non Probability Sampling

• When random samples are not possible

• Rare disease
• Small population
• Special population
• Special Condition
• Difficult to reach population

12/08/2012 Dr. Kusum Gaur 133

 Non-probability Samples

Convenience
 Purposive
 Quota
 Snow ball study

12/08/2012 Dr. Kusum Gaur 134

12/08/2012 Dr. Kusum Gaur 135
12/08/2012 Dr. Kusum Gaur 136
12/08/2012 Dr. Kusum Gaur 137
 Snow ball sampling

Contact tracing
Initial respondent helps in recruiting
new population
Useful in network analysis approach

12/08/2012 Dr. Kusum Gaur 138

 Step-4 & 5

Data Collection
and
Data Management
 Sources of Data

• Primary –Own generated data

• Secondary –Already generated data

Published
Non-Published

12/08/2012 Dr. Kusum Gaur 140

 Primary Vs Secondary source of Data

Primary data Secondary data

• Need to be generated • Readily available

• First hand information • Second hand information

• Questionnaire
• Not need of questionnaire
• Purpose served
• Purpose served ?
• Analysis as per
purpose • Descriptive
• Require more time and
• Less expensive
money
12/08/2012 Dr. Kusum Gaur 141
Type of Data Collection Methods

 Interview
Personnel
Telephonic
 Observation
 Experimental
 Interview and Observation
 Observation and Experimental
 Interview ,Observation and Experimental
 

12/08/2012 Dr. Kusum Gaur 142

Forms of questions(Open Vs Closed)

Open ended Close ended

• Possible responses are • Categories are given
not given. already coded
• Mean, SD, Median • Proportion
• For seeking opinions, • For eliciting factual
attitudes ,perceptions information
• Not so depth
• Provides in depth info.
• Investigator’s bias
• Experience of • Ease of answering,
investigator and • Easy to analyse
analyst required

12/08/2012 Dr. Kusum Gaur 143

Considerations in formulating questionnaire

(Questionnaire/Interview schedule)

 Use simple and everyday language

 Do not use ambiguous questions(?/?)

 Do not ask leading questions

 The order of questions:

 Guideline for filling an instrument, pen-pencil

Pre testing

12/08/2012 Dr. Kusum Gaur 144

 Validity of a Research Instrument

Ability of an instrument to measure what it is

designed to measure being measured

Establish the logical link between the

questions and objectives

 Items/questions cover the full range of

issue/attitude being measured

12/08/2012 Dr. Kusum Gaur 145

1.Decide the information required. Steps
2. Define the target respondents.

3. Method(s) of reaching target

4. Decide on question content.

5. Develop the question wording.

6. Put questions into a meaningful order.

7. Check the length of the questionnaire.

8. Pre-test the questionnaire.

9. Develop the final survey form

12/08/2012
Steps of Questionnaire Design Dr. Kusum Gaur 146
 Organization and Compilation of Data

Organization and Compilation of Data in such a way

(Master Chart ) to have reliable, relevant, adequate
and reasonably complete data with following
requisites –
Simplicity
Briefness
Utility
Distinctively
Comparability
Scientific Arrangement
Attractive
Effective
12/08/2012 Dr. Kusum Gaur 147
Observations
 Steps of Observations

• Entry of Observations Unites

• Master Chart

• Tabulation

• Diagrammatic Presentation

12/08/2012 Dr. Kusum Gaur 149

Entry of Observation Unites
Master Chart
Grouping & Classification
Grouping & Classification
Grouping & Classification
Grouping & Classification
Grouping & Classification
Grouping & Classification
Grouping & Classification
Grouping & Classification
Grouping & Classification
Grouping & Classification
Grouping & Classification
Grouping & Classification
Grouping & Classification
Grouping & Classification
Grouping & Classification
Master Chart for Analysis
 Tabulation – Content of Table
 Table No. Sequence in the text
 Tile of Table –short, clear and self explanatory to say about for what the
table is ?
 Body of Table –consist of rows and columns
 Rows – 1st row shows headings of columns
 1st column shows headings of rows
 rest of rows and columns are showing data as per required
 number of rows and columns should be limited to maintained
simplicity of table
 source of data ( if it is other than the present study ) should be written
just below the body of table
 Source of Data ?
 Foot Note - written just below the body of table, if there is any hidden
information
 Inferences –summary value of table

12/08/2012 Dr. Kusum Gaur 168

 Types of Tables


As per purpose
General tables –about Socio-demographic profile
Specific tables –about Aims and objectives


As per originality
Original tables-from original Data
Derived tables –from original tables


As per Construction
Simple tables- showing one variable at one time
Complex tables – showing > one variable at one time

12/08/2012 Dr. Kusum Gaur 169

12/08/2012 Dr. Kusum Gaur 170
Tabulation
 Diagrammatic Presentations
Bar
Qualitative
 Simple Data
Histogram
 Multiple Frequency Polygon
 Component Cumulative Frequency
Pie Polygon
Line Scatter Diagram
Pictogram Box and Whisker
Spot Map Correlation Diagram

Qualitative Data Quantitative Data

12/08/2012 Dr. Kusum Gaur 172
12/08/2012 Dr. Kusum Gaur 173
Diagrammatic Presentations
 Simple Bar diagram

12%
4th Qtr

3rd Qtr
14%
12%

2nd Qtr
32%

1st Qtr
82%

0 1 2 3 4 5 6 7 8 9

12/08/2012 Dr. Kusum Gaur 175

 Multiple Bar diagram
60

50

40

(1) 1-5 Years

30
(2) 6-10 Years
(3) 11 & Above Years

20

10

0
(1) Very Dissatisfied (2) Dissatisfied (3) neither satisfied (4) Satisfied (5) Very Satisfied
nor dissatisfied

12/08/2012 Dr. Kusum Gaur 176

 Component Bar diagram

12/08/2012 Dr. Kusum Gaur 177

 Pie diagram

Propotion of Pie = (Proportion of that variable )(360)Degree

12%

14% 1st Qtr

2nd Qtr
82% 3rd Qtr
4th Qtr
32%

12/08/2012 Dr. Kusum Gaur 178

 Line diagram
7

4
Series 2
3
Series 1
2

0
2000 2001 2002 2003 2004 2005

12/08/2012 Dr. Kusum Gaur 179

 Histogram ( Area Diagramme)

Series 1
40
30
20
10
Series 1
0
0 to 5 yrs
5yrs to 10
10 yrs to
yrs 15 yrs to
15 yrs 20 yrs to
20 yrs
25 yrs

12/08/2012 Dr. Kusum Gaur 180

 Scatter Diagram
30

25

20
Duration of Diabetes

15
Duration of diabetes in yrs.
Linear (Duration of diabetes in yrs.)

10

0
0 50 100 150 200 250 300

No. of Patients

12/08/2012 Dr. Kusum Gaur 181

 Radar diagram

5/1/2002
40
30
20
9/1/2002 6/1/2002
10
Series 1
0
Series 2

8/1/2002 7/1/2002

12/08/2012 Dr. Kusum Gaur 182

 Box & Whisker
70

60

50

40 Open
High
30 Low

20 Close

10

0
5/1/2002 6/1/2002 7/1/2002 8/1/2002 9/1/2002

12/08/2012 Dr. Kusum Gaur 183

 Step-6

Analysis of Data
 Biostatistics = Biology + Statistics

• Biostatistics is application of statistics in

biology i.e. science of figure in medical science

• Data: Set of information, facts or figures

numerically coded and from which conclusions
may be drawn is called data (singular-datum).

• Statistics: The collection of methods used in

planning an experiment
and analyzing data in order to draw accurate
conclusions.
 Type of Biostatistics

• Descriptive statistics generally characterizes

or describes a set of data elements

• Inferential statistics tries to infer information

about a population by using information
gathered by sampling
 
 Descriptive Analysis

Qualitative Data
Rates
Ratios
Proportions

Quantitative Data
Central Tendencies  Disperson
Mean Standard Deviation
Mode Standard Error
Median Confidencial Limit
Skeweness

12/08/2012 Dr. Kusum Gaur 187

 Descriptive Analysis of
Qualitative Data
No. of total Events in a year (A)
Rate = * 1000
MYP of that Region (T)

No. of total (A)

Ratio =
No. of total (B)
No. of Specific Events (A)
Percentage of Events = * 100
Total Events (T)

Event of Sp. Cause (A)

Proportional Rate = * 10 n
Total Deaths (T)
12/08/2012 Dr. Kusum Gaur 188
 Descriptive Analysis of
Quantitative Data
Mean = Mathematical Average ∑X
N
Mode = Most commonly occurring value
Median = Center value when arrange in increasing N+1
or decreasing fashion 2

Standard Deviation = It tells how much scores deviate from the mean
 it is the square root of the variance (X-X)
 it is the most commonly used measure of spread SD=√ N

Standard Error = Deviation from mean per observation

SD/ √N

Skewness = Deviation of peak from median

SK= 3 (Mean –Median)/SD
12/08/2012 Dr. Kusum Gaur 189
SD from MS Excel
SD from MS Excel…
 Appropriate choice

of

significance tests

12/08/2012 Dr. Kusum Gaur 192

 TEST OF SIGNIFICANCE OF QUALITATIVE DATA

 
TEST OF SIGNIFICANCE OF QUALITATIVE DATA
 
 
One Sample Two Sample >Two Sample
 
Sample proportion
to Independent Dependent Dependent Independent
Population Proportion
Mc Numer Cochron’s

Large Sample Small Sample

(>30) (<30) Yat’s Corrected
‘Z’ Score
Small Corrected
Sample Large Sample‘Z’ Score
Large Sample Small Sample
Chi Squire
Chi Squire ‘Z’ Score Chi Squire
Yat’s Corrected Chi
Chi Squire

12/08/2012 Dr. Kusum Gaur 193

TEST OF SIGNIFICANCE OF QUANTITATIVE DATA

 
TEST OF SIGNIFICANCE OF QUANTITATIVE DATA
 
 
One Sample Two Sample >Two Sample
 
Sample Mean
to Independent Dependent Dependent Independent
Population Mean
Paired ‘T’ Test ANOVA Friedman

Large Sample Small Sample

(>30) (<30)
‘Z’ ‘T’ Test
TestSample Large Sample
Small Large Sample Small Sample
‘Z’ Test
ANOVA ANOVA

12/08/2012 Dr. Kusum Gaur 194

 STUDY DESIGNS AND APPROPRIATE TEST

Type Study Design Appropriate Significance Test

Descriptive Study

Analytical
Case Control Study OR
Qualitative ‘Z’ Score Test/Chi-Square Test
Quantitative ‘Z’ Test/’t’ Test
Cohort study OR, AR, & RR
Qualitative ‘Z’ Score Test/Chi-Square Test
Quantitative ‘Z’ Test/’t’ Test
12/08/2012 Dr. Kusum Gaur 195
 STUDY DESIGNS AND APPROPRIATE TEST
Type Study Design Appropriate Significance Test

Randomized Controlled study

Quantitative (before and after)- Paired ‘t’ Test
Quantitative (before and after >1 followup)- Freidmen ANOVA
Quantitative (between two Gps)- Unpaired ‘t’ Test
Quantitative (between > two Gps)- ANOVA Test

Randomized Controlled study

Qualitative (before and after)- Mac Numer Test
Qualitative (before and after >1 followup)- Cochron’s Test
Qualitative (between two Gps)- ‘Z’ Score/Chi-square Test
Qualitative (between > twoDr.Gps)-
12/08/2012 Chi-square Test
Kusum Gaur 196
 STATISTICAL TEST OF
SIGNIFICANCE
Nominal Numerical Ordinal

Two Groups ‘Z’ Score Test ‘Z’ test (n>30) Mann Whitny
Chi-square Test T Test (n<30)

> Two Groups Chi-square Test ANOVA Kruskal Wallis

Paired Two Mec Numer Paired ANOVA Wilcoxon Sign

Multiple Cohrane Repeated Friedman

Observation in Multivarient ANOVA
same individual

Association of Contegency Correlation(Pearson) Spearman

Two Variable Cofficient Regression Correlation
 STATISTICAL TEST OF SIGNIFICANCE
Research Number and Number and Covariates Test Goal of Analysis
Question type of DV type of IV

Nominal 1 nominal chi square determine if difference between

Group croups
differences
Continuous 1 dichotomous t-test
Determine significance of
1 Categorical 1 one-way ANOVA mean group
1+ one-way differences

ANCOVA
2+ Categorical 1 factorial ANOVA
1+ factorial ANCOVA
2+ Continuous 1 Categorical 1 one-way MANOVA Create linear
1+ one-way MANCOVA combo of Dependent variable
2+ Categorical 1 factorial (Dvs)
MANOVA to maximize
1+ factorial MANCOVA mean group
differences
Degree of Continuous 1 Continuous Bivariate Determine
relationship Correlation relationship/prediction

2+ Continuous Multiple Linear combination to predict the

Regression DV
1+ Continuous 2+ Continuous Path Analysis Estimate causal relations among
variables

12/08/2012 Dr. Kusum Gaur 198

 Comparing difference between
Two Sample Proportions
‘Z’ Score Test
P2 – P1 here, P1– proportion of that event in 1st Sample
‘Z’ Score = P2 - proportion of that event in 2nd Sample
SEDP SEDP – Standard Error of
Difference in Proportion

Q1 - proportion without that event

in 1st Sample i.e. Q1 = 100 – P1
P1Q 1 P2Q2 Q2 - proportion without that event in
SEDP = ------- + -------- 2nd Sample i.e. 100 – P2
N1 N2 N1 - Sample Size of 1st Sample
N2 - Sample Size of 2nd Sample

12/08/2012 Dr. Kusum Gaur 199

 Inference of ‘Z’ Score Test

If ‘Z’ > 2 = Difference is Significant

If ‘Z’ < 2 = Difference is Not Significant

If ‘Z’ > 3 = Difference is Highly Significant

12/08/2012 Dr. Kusum Gaur 200

 Comparing difference between
>Two Sample Proportions
Chi-Square Test
 
Indications
Qualitative data
Normal distribution
Comparing difference between
Two Sample proportions
Multiple Sample proportions

12/08/2012 Dr. Kusum Gaur 201

 Comparing difference between
>Two Sample Proportions
Chi-Square Test
Chi Square(2) = ∑all cells(O-E)2 Tr x T c
E E=
T
(O1-E1)2 (O2-E2)2 (O3-E4)2 (On-En)2
Chi Squire = + + + ---+
E1 E2 E3 En
Tr – Total of that Row
here, O – Observed value of cell Tc – Total of
that column
E – Expected value of cell,
considering Null T – Grand Total i.e. a+b+c+d
Hypothesis
Degree of Freedom (DF) = (C – 1) (R -1)

12/08/2012 Dr. Kusum Gaur

R= No. of Rows, C = No. of Column 202
Inference of Chi Square(x2)
Chi Square(x2 ) value is seen at Degree of Freedom
DF = (R – 1) (C – 1), from Chi Square((2) Table
(here R=No. of Rows &C= No. of Column)
at desired level of significance

Inferences
If Chi Square(x2 ) Test Value is –
Higher than Table value = Difference in proportions is
Significant at that desired level of significance.

If Chi Square(x2 ) Test Value is –

Lower than Table value = Difference in proportions is
Not Significant at that desired level of significance.
12/08/2012 Dr. Kusum Gaur 203
 Comparing difference between
Two Sample Means (>30)
‘Z’ Test
Pre-requisites
Quantitative data
 Homogenous normally distributed Random Sample
Sample Size > 30
 
Indications
To see the Significance of any Observation in reference of
Mean Value of that sample
Comparing difference between
Sample Mean to Population Mean
Means of Two independent Samples

12/08/2012 Dr. Kusum Gaur 204

 Comparing difference between
Two Sample Means (>30)
‘Z’ Test
X2 – X1 here, X1– Mean of that event in 1st Sample
‘Z’ Test = X2 - Mean of that event in 2nd Sample
SEDM SEDM – Standard Error of
Difference in Means

SD1 – Standard Error of 1st Sample

SD2 – Standard Error of 2nd Sample
SD2 1 SD2 2 N1 - Sample Size of 1st Sample
SEDM = ------- + -------- N2 - Sample Size of 2nd Sample
N1 N2

12/08/2012 Dr. Kusum Gaur 205

 Comparing difference between
Two Sample Means (<30)
‘T’ Test
Prerequisites

Random Sample

Quantitative data

Normally Distributed

Sample Size < 30

12/08/2012 Dr. Kusum Gaur 206
 Type of ‘T’ Test

as per design
Unpaired / Paired

for inference
One Tail /Two tail

12/08/2012 Dr. Kusum Gaur 207

 Unpaired ‘T’ Test Design

Population -1 Population -2

S-1 S-2

Mean --1 Unpaired ‘T’ test Mean --2

12/08/2012 Dr. Kusum Gaur 208

 Paired ‘T’ Test Design

Intervention

Population Sam
Observations-1 Observations 2
ple-

Mean --1 Mean --2

Paired ‘T’ test

12/08/2012 Dr. Kusum Gaur 209

 One Tail ‘T’ Test

Acceptance Zone Rejection Zone

One Tail – Results are aspect only in one direction
 Two Tail ‘T’ Test

Rejection Zone Acceptance Zone Rejection Zone

Two Tail – Results are aspect in both direction
 Comparing difference between
Two Sample Means (<30)
‘T’ Test
X2 – X1 here, X1– Mean of that event in 1st Sample
‘T’ Test = --------------- X2 - Mean of that event in 2nd Sample
SEDM SEDM – Standard Error of
Difference in Means

SD1 – Standard Error of 1st Sample

SD2 – Standard Error of 2nd Sample
SD2 1 SD2 2 N1 - Sample Size of 1st Sample
SEDM = ------- + -------- N2 - Sample Size of 2nd Sample
N1 N2

Degree of Freedom (DF) = (N1 – 1) + (N2 -1) = N1 + N2 - 2

12/08/2012 Dr. Kusum Gaur 212

Inference of ‘T’ Test Value
‘T’ Test Value is matched at Degree of Freedom
(DF) = N1 + N2 – 2 in the Table of “T”
at desired level of significance.

Inferences
If ‘T’ Test Value is –
Higher than Table value = Difference in Means is
Significant at that desired level of significance.

If ‘T’ Test Value is –

Lower than Table value = Difference in Means is
Not Significant at that desired level of significance.
12/08/2012 Dr. Kusum Gaur 213
 Comparing difference between
>Two Sample Means

ANALYSIS OF VARIENCE (ANOVA) TEST

  Pre-requisites
Quantitative data
 Homogenous normally distributed Random Sample
 
Indications
Comparing difference between more than Two
Means

12/08/2012 Dr. Kusum Gaur 214

 Comparing difference between
>Two Sample Means
‘ANOVA’ Test
MSOSI MSOS2 - Mean Sum Of Squares Within Classes
ANOVA = ---------- = Total SOS – MSOSI
MSOS2
T SOS = X2 – (X)2/N

MSOSI – Mean Sum Of Squares Between Classes = SOSI / K-1

SOSI –Sum Of Squares Between Classes

(Xa)2 (Xb)2 (Xc)2 (Xk)2 (X)2

= --------- + ----------- + ----------- + ….+ ____ __ - --------- Na Nb Nc
Nk N

At Degree of Freedom (DF) = ( K-1) Horizontal

(N – K) Vertical
12/08/2012 Dr. Kusum Gaur 215
Inference of ANOVA
Find out Variance Ratio value at Degree of Freedom
(DF) = ( K-1) Horizontal, (N – K) Vertical
from the Variance Ratio Table
at desired level of significance.

Inferences
If Test value is > Table value = Difference in Means is
Significant at that desired level of significance.

If Test value is < Table value = Difference in Means is

Not Significant at that desired level of significance.

12/08/2012 Dr. Kusum Gaur 216

 CORRELATION

Indications

To find out relationship between variables

12/08/2012 Dr. Kusum Gaur 217

 Type & Degree of Correlation
Correlation Inference Correlation (r) Inference
+1 Perfect +ve -1 Perfect +ve
Correlation Correlation
> 0.95 About Perfect +ve > - 0.95 About Perfect +ve
Correlation Correlation
> 0.75 V. Good Correlation > - 0.75 V. Good Correlation

0.75 – 0.5 Moderate Correlation - 0.75 to – 0.5 Moderate

Correlation
0.5 – 0.25 Fair Correlation - 0.5 to – 0.25 Fair Correlation
0.25 - 0 No Correlation < - 0.25 No Correlation

12/08/2012 Dr. Kusum Gaur 218

 Correlation

CORRELATION
 

Two Variables > Two Variables

 

Un-Paired Data Paired Data 

Pearson’s Spearman’s Rank Order Multivariate

Correlation Correlation Correlation
 

12/08/2012 Dr. Kusum Gaur 219

 Pearson’s correlation

. ∑ ( X – X) ∑ ( Y – Y) ∑xy
Correlation (r) = =
√∑ ( X – X)2 ∑ ( Y – Y)2 √ ∑ x2 y2
 
 
Direct Method
∑ X Y - ∑ X ∑Y / N
Correlation (r) = -----------------------------
√ {∑X2 – (∑X)2/N}{ ∑Y2 – (∑Y)2 /N}

12/08/2012 Dr. Kusum Gaur 220

 Pearson’s correlation -----

 
here,
∑ X Y = Sum of multiplication of X and Y
∑ X = Sum of all observations of X Series
∑ Y = Sum of all observations of YX Series
N =Total no. of observations
∑X2 = Sum of Squares of all observations of X Series
∑Y2 = Sum of Squares of all observations of Y Series
(∑X)2 = Square of Sum of all observations of X Series
(∑Y)2 = Square of Sum of all observations of Y Series

12/08/2012 Dr. Kusum Gaur 221

 Spearman’s Rank Order Correlation

6∑D2
• Spearman’s Rank (rs ) = 1 -
N3 - N
 
 

12/08/2012 Dr. Kusum Gaur 222

 Significance Test for Correlation (r)

Standard Error (SE) of rs = rs √ N-1

Inference
• If difference >2 SE of r =Difference is
Significant at 5% level
• If difference < 2SE of r =Difference is
Not Significant at 5% level

12/08/2012 Dr. Kusum Gaur 223

 REGRESSION

Indication
To find out causal relationship between
variables

REGRESSION COFFICIENT- It is a measure of

change in one dependent variable (y) with
one unit change in the other variable (x)

12/08/2012 Dr. Kusum Gaur 224

Regression line with Regression Equation

The regression equation of ‘Y’ on ‘X’ is expressed as follows:

Here, ‘a’ is interceptor & ‘b’ is slope Yc = a + bX
 Regression Lines

Régression line of Y on X is Y = a + bX ----(1)

Régression line of X on Y is X = a + bY ----(2)

Here- Y = one variable

X = other variable
a = interceptor of X line on Y line
b = slope of X line on Y line Regression

12/08/2012 Dr. Kusum Gaur 226

 Regression – Equations
Regression Equation of X on Y
SD of series X
(X – X)= r (Y –Y) ---- (3)
SD of series Y
 

Regression Equation of Y on X

SD of series Y
(Y – Y)= r (X –X) ------- (4)
SD of series X
 

12/08/2012 Dr. Kusum Gaur 227

 Regression – coefficients
Regression Coefficient of X on Y
SD of series X ∑(X-X)(Y –Y)

b(xy)= r =
SD of series Y ∑(X – X)2
 

Regression Coefficient of Y on X
SD of series Y ∑(X-X)(Y –Y)
b(yx)= r =
SD of series X ∑(Y – Y)2

12/08/2012 Dr. Kusum Gaur 228

 Relation of correlation and
Regression

Co-rrelation (r) = √ bxy byx

 

12/08/2012 Dr. Kusum Gaur 229

 Between
Tests/Procedure/Therapy
For comparison with Gold Standard:
Sensitivity
Specificity
PPV
NPV
ROC

For agreement of association: Kappa

For appropriate cut of value for diagnostic test: ROC

12/08/2012 Dr. Kusum Gaur 230

 Sensitivity and Specificity
    Status based on gold standard test

    Diseased Normal

Test positive True positive False positive

Observation in a b
new test Test negative False negative True negative
c d

Sensitivity = a /(a+c) PPV = a /(a+b)

Specificity = d /(b+d) NPV = d /(c+d)

12/08/2012 Dr. Kusum Gaur 231

‘ROC’ Curve
 Kappa Statistics
(Measurement of Agreement)
Test Value Inference
0.93 – 1 Excellent Agreement
0.81 – 0.92 Very Good Agreement
0.61 – 0.80 Good Agreement
0.41 – 0.60 Fair Agreement
0.21 – 0.40 Slight Agreement
0.01 – 0.20 Poor Agreement
< 0.01 No Agreement
12/08/2012 Dr. Kusum Gaur 233
 Non-Parametric Tests
Advantages
Distribution free
Easier to do
Easier to understand/infer

Disadvantages
They ignore certain amount of information
Indicated only ordinal or nominal data
Statistically Less efficient
Indicated only to test hypothesis, not for estimates

12/08/2012 Dr. Kusum Gaur 234

 Parametric Test Vs Non-Parametric
Test Quality Parametric Non-Parametric

Assumed Distribution Normal Any

Assumed Variance Homogenous Any

Data Type Interval-Continous Nominal /Ordinal

Data set Relationship Independent Any

Usual Centre Measure Mean Median

More conclusions Easier to calculate

Advantages
More efficient Less affected by outliers

12/08/2012 Dr. Kusum Gaur 235

Parametric Test Vs Non-Parametric
Test Parametric Non-Parametric

Correlation test Pearson Spearman

Independent Independent-
Mann-Whitney test
measures, 2 groups measures t-test

One-way,
Independent
independent- Kruskal-Wallis test
measures, >2 groups
measures ANOVA

Repeated measures,
Matched-pair t-test Wilcoxon test
2 conditions

Repeated Signmeasures, One-way, repeated

Test (K Test)– nonparametric test for quantitative paired data
Friedman's test
>2 conditions
12/08/2012 measures ANOVA
Dr. Kusum Gaur 236
 Sign test

• Simplest
• Based on direction(- /+/0)
• Signs as per the direction are counted

• Inference – if S≤K = Null hypothesis (H₀) is

rejected
• Here ‘S’ is net sum of signs as per sign
• ‘K’ is constant

12/08/2012 Dr. Kusum Gaur 237

 Sign test – Steps

Sign K Test for Small Sample (<30)

– Find out net sum of signs as per sign(S)
– S = (total + signs) – (total – signs)
– K = (n-1)/2 - 0.98√n
• Inference – if S≤K = Null hypothesis (H₀) is rejected

Sign Z Test for Large Sample (>30)

– Find out no of ties with less frequent sign(X)
– Z = (X – np) / √ np (1-p) here X= no. + Sign
• Inference – if Z>2 = Null hypothesis is rejected

12/08/2012 Dr. Kusum Gaur 238

12/08/2012 Dr. Kusum Gaur 239
12/08/2012 Dr. Kusum Gaur 240
12/08/2012 Dr. Kusum Gaur 241
12/08/2012 Dr. Kusum Gaur 242
12/08/2012 Dr. Kusum Gaur 243
 Step-7

Inferences

12/08/2012 Dr. Kusum Gaur 244

 Steps in Statistical Inference

 Generating NULL and ALTERNATIVE

hypothesis
 Testing the hypothesis using appropriate
statistical tests
 Obtaining ‘p’ value
 Concluding from the p value.
 Obtaining Level of Significance
 Comparing ‘p’ value with CI.

12/08/2012 Dr. Kusum Gaur 245

 ‘P’ Value and Inferences
with Normal Curve

12/08/2012 Dr. Kusum Gaur 246

 Rejection Zone Acceptance Zone Rejection Zone
Mean ± 1SD =68% values - Confidence Limit 68% - P Value = >0.05 - NS
Mean ± 2SD =95% values - Confidence Limit 95% - P Value = 0.05 - S
Mean ± 3SD =99% values - Confidence Limit 99% - P Value = 0.001 - HS
 Normal Curve, P Value and Level of Significance

Rejection Zone Acceptance Zone Rejection Zone

Mean ± 1SD =68% values - Confidence Limit 68% - P Value =/>0.05 - NS
Mean ± 2SD =95% values - Confidence Limit 95% - P Value < 0.05 – S
Mean ± 3SD =99% values - Confidence Limit 99% P Value < 0.001 - HS
12/08/2012 Dr. Kusum Gaur 248
 Conventionally Accepted
Significance Level

 P Value > 0.05 LS=Not Significant

 P Value < 0.05 LS=Significant

 P Value < 0.001 LS=Highly Significant

 Step-8

Reporting

12/08/2012 Dr. Kusum Gaur 250

 Steps of Report Writing

Title of Project
Abstract
Introduction
Aims & Objectives
Methodology
Observations-Compilation, Classification &
Presentation of data with analysis and inferences
Discussion
Conclusions
Recommendations
Limitations
Acknowledgment
Bibliography

12/08/2012 Dr. Kusum Gaur 251

 Discussion

 Explanation of findings
 Logic and reasoning for the results as it
appears
 Compare and contrast with findings of other
researchers
 Based on objectives of the study
 Should answer the research question
 Scope & limitations of the study

12/08/2012 Dr. Kusum Gaur 252

 Recommendations & conclusions

• Based on our findings

• Limited to objectives of the study
• Policy implications
• Relevance should be emphasized
• Should be exclusively limited to
observations

12/08/2012 Dr. Kusum Gaur 253

 Managerial and financial aspects

 Protocol development
 Time line/Gantt chart
 Peer review
 Development of tools
 Training in data collection
 Budget/ financial accounting
 Quality control
 Monitoring & Evaluation

12/08/2012 Dr. Kusum Gaur 254

 Time Line/Gant chart/log Fram
Activities 1.1.12- 16.1- 1.2.12- 1.3.12- 16.5.12- 16.6.12- 16.7.12-
15.1.12 31.1 15.2.12 15.5.12 15.6.12 15.7.12 31.7.12

Planning
Officials
Que. Dev
Training
Poilet Survey
Corrections
Re-training
Resource Proc

Survey
Analysis
Report Writing
Dissemination
of Report
 Computer in Statistics

12/08/2012 Dr. Kusum Gaur 256

 Web sites related to Statistics

• http://stattrek.com
• http://vassarstat.net
• http://www.scribd.com
• http://www.statistixl.com
• http://statistics calculators.com
• http://stat.ubc.ca/~rollin/stats/ssize/
• ………………………………………………………
……
12/08/2012 Dr. Kusum Gaur 257
 Computer Softwares in Statistics

• Microsoft Excel
• SPSS
• Epi info
• Epi tab
• Mini tab
• Graph Pad
• Primer
• Medcal
• ……………..
12/08/2012 Dr. Kusum Gaur 258
 Always there is room for improvement

12/08/2012 Dr. Kusum Gaur 259