Module 4:

Data analysis and presentation

Six steps in the IR process
 Quantitative vs
Qualitative

What are the differences between

quantitative and qualitative research?
Research questions
Methodological differences
Data analysis
 Comparing qualitative and quantitative approaches

Qualitative Quantitative
Social theory Action Structure
Methods Observation, interview Experiment, survey
What is x? How? Why? How many xs?
Question
(classification) (enumeration)
Reasoning Inductive Deductive
Sampling Theoretical Statistical
Strength Validity Reliability

Pope and Mays (1995). Reaching the parts other methods cannot reach: an introduction to qualitative methods in health and health
services research. BMJ: 311; No. 6996
 Module 4a:
Quantitative data analysis and
presentation
 Presentation outline

Expected outcomes
Key concepts
Data analysis plan
Quantitative data analysis
Data management
 Learning Objectives &
Expected outcomes
Able to:
Describe data analysis planning processes
Understand appropriate statistical measures
Understand data management approaches
Appreciate the importance of tailored / audience
sensitive data presentation
 Key concept 1:
Data analysis plan

Designing analysis for use IR aims to:

Understand the implementation processes
Communicate the implementation process
to stakeholders

“Emphasis on simplicity and interpretability”

 Key concept 1:
Data analysis plan
Designing analysis for use
Different stakeholders need different information:
Lay people?
Community leaders?
Local government/health service leaders?
Civil society and media personnel?
National policy-makers?
“Emphasis on simplicity and interpretability”
 Key concept 1:
Data analysis plan

Designing analysis by purpose focuses

on the objective of the analysis:
Effectiveness
Efficiency
Equity
Sustainability
 Key concept 1:
Data analysis plan
Data presentation formats
Data reporting should be presented in both
textual and visual formats, such as:
Tables
Diagrams
Graphs
Infographics
Maps
Provider education expressed as frequency table

Level of education of private providers Frequency

Illiterate 106

Basic literacy 74

Primary school certificate 57

Secondary school certificate 11

Higher level qualification 2

Total 250
Joint frequency distributions for two or more variables

Highest level Men Women All

Illiterate 42 64 106
Basic literacy 45 29 74
Primary school certificate 32 25 57
Secondary school certificate 8 3 11
Higher level qualification 1 1 2
Total 128 122 250
 Provider education presented as proportion, percentage
and cumulative %

Cumulative
Level of education Proportion Percentage
percentage
Illiterate 0.424 42.4 42.4
Basic literacy 0.296 29.6 72.0
Primary school
0.228 22.8 94.8
certificate
Secondary school
0.044 4.4 99.2
certificate
Higher level
0.008 0.8 100.0
qualification
Total 1.000 100.0
 Row percentages

Highest level Men Women All

Illiterate 39.6 60.4 100.0
Basic literacy 60.8 39.2 100.0
Primary school certificate 56.1 43.9 100.0
Secondary school certificate 72.7 27.3 100.0
Higher level qualification 50.0 50.0 100.0
Total 51.2 48.8 100.0
 Column percentages

Highest level Men Women All

Illiterate 32.8 52.5 42.4
Basic literacy 35.2 23.8 29.6
Primary school certificate 25.0 20.5 22.8
Secondary school certificate 6.3 2.5 4.4
Higher level qualification 0.8 0.8 0.8
Total 100.0 100.0 100.0
Provider education expressed as a bar chart
Bar chart for two variables
Provider education presented as a pie chart
 Line graph for trend analysis

Average Ae. Aegypti population per week in 5 field study sites

In what other format(s)
can this data set be
presented?
Map for spatial distribution
20th century death

Source: Information is beautiful

url: http://www.informationisbeautiful.net/visualizations/20th-century-death/
Potential tax revenue from drugs

Source: Information is beautiful site

http://www.informationisbeautiful.net/visualizations/
drug-deal-potential-tax-revenue-from-legalized-narcotics/
Photo
Interactive graph

Source: Gapminder
http://www.gapminder.org/world-offline/
Whiteboard animation

Source: Eliminate Dengue Program

URL: http://www.eliminatedengue.com
 Reflection activity

In your project, discuss the results of the study

that you need to disseminate and format of data
presentation you will use for different
stakeholders.
 Key concept 2:
Quantitative data analysis

Depending on research question:

Descriptive vs analytic study?
Analytic study, what to find?
Association
Causality
Statistical difference
 Key concept 2:
Quantitative data analysis
Variables in quantitative analysis
are usually classified by their level
of measurement:
Rational
Interval
Ordinal
Nominal
 Key concept 2:
Quantitative data analysis

Descriptive statistics
Distributions and summary measures
Defining intervals for frequency distributions
Frequency distribution and summary statistics
Measures of variation
 Key concept 2:
Quantitative data analysis

Distributions and summary measures

Advantages of frequency distributions:
useful for all types of variables
easy to explain and interpret
presented graphically and in different formats
 Key concept 2:
Quantitative data analysis
Defining intervals for frequency
distributions
Constructing a frequency distribution requires a choice
of intervals:
Ordinal
Interval
Rational
Two conflicting objectives when determining intervals:
Limiting the loss of information
Providing a simple, interpretable and useful summary
 Key concept 2:
Quantitative data analysis

Summary statistics and frequency

distributions
A powerful and robust form of analysis. Summary
statistics usually focus on: overall location of a
distribution or extent of variation within a
population.
 Key concept 2:
Quantitative data analysis

Use of mean or median

Mean – the average value
Median – the value in the middle
 Use of mean or median

Normal distribution Skewed distribution

 Key concept 2:
Quantitative data analysis

Measures of variation
How much variability?
Low variability
High variability
 Key concept 2:
Quantitative data analysis
Measures of variation
Choices of measures
Variances
Standard deviations

Alternative measures
Quartiles: divide data into four quarters
(Q1 to Q4) – 25% in each
Percentiles: divide the data into two parts
 Key concept 2:
Quantitative data analysis

Analytical statistics
Group comparison
Association
Causality
 Key concept 2:
Quantitative data analysis
Measurement Assumption of Type of group Analysis
scale distribution
Nominal or - Independent Chi square test
Ordinal
- Paired Sign test
Interval or Normally Independent Independent test
Ratio distributed
Paired Paired test
Not normally Independent Median test
distributed
Paired Wilcoxon
 Key concept 2:
Quantitative data analysis
Finding association
Pearson correlation
Ratio/interval scale
Normal distribution of data
Rank correlation
Ratio/interval scale
Non normal distribution of data
Chi Square
Categorical data
 Key concept 2:
Quantitative data analysis

Causality (regression)
Linear regression
Continuous variable of both independent and
dependent variable
Normal distribution of data
Logistic regression
Dichotomous dependent variable
Continuous and categorical independent variable
 Key concept 2:
Quantitative data analysis

Causality (regression)
Cox proportional hazard model
Time-dependent outcome (survival model)
Continuous and categorical independent variable
 Key concept 2:
Quantitative data analysis

Measures of risk
‘Risk’ and ‘odds’ used interchangeably, but
not the same
‘Reduction in risk’ is not equivalent to
‘reduction in odds’
 Key concept 2:
Quantitative data analysis

Measures of risk:
The ‘denominator problem’
Risk calculation requires calculation of the
population at risk
Provide the estimates of both the numerator and
denominator alongside any proportion,
percentage or risk estimate
 Key concept 2:
Quantitative data analysis

Sub-group analysis
The outcomes of an intervention may differ
among sub-groups.
‘Data mining’ is useful to formulate new
hypotheses but requires great caution in IR.
 Reflection activity

In your project, discuss the data analysis that

you will do and identify whether the data you
are collecting is suitable for the type of
analysis you plan.
 Key concept 3:
Data management

Principle of data management

Data management and study phase
 Key concept 3:
Data management

Data quality and integrity

Data should be:
High quality
Reliable
“No study is better than the quality of its data”
 Key concept 3:
Data management
Prior to data collection process
ID number
Flow of data collection and handling process
Protocol for quality control
Checking interviewee response
Re-interview process
Electronic database development
SOP for data entry process
 Key concept 3:
Data management

Data collection process

Data collection supervision
Questionnaires/data collection forms storage
management
Checking data entry process
 Key concept 3:
Data management

Post-data collection process

Checking database consistency
Data cleaning
Data coding
 Reflection activity

In your project proposal, discuss how to

improve the quality of their data
management system.
 Conclusion
Start from the end
Plan your data analysis according to
stakeholders’ need for information
Use appropriate statistical tools
according to the information needed
Manage your data to ensure the validity
of the data collected