The document provides an overview of data analysis, interpretation, and presentation for research. It discusses quantitative and qualitative data analysis methods. Quantitative methods include descriptive statistics like measures of central tendency and spread to summarize numeric data. Qualitative methods include coding of open-ended responses to identify themes. Both types of data can be presented visually in tables, graphs, and word clouds. Proper interpretation relates results back to the research question and previous studies and draws conclusions about the meaning and implications of the findings.

1. RESEARCH 101 TRAINING
How to be Smart About Research in 8 Hours or Less
Session 3: Data analysis,
interpretation, and presentation
July 13, 2017 @ 5:30-7:00 p.m. Eastern

2. Copyright 2017 Global Healthy Living Foundation, Inc.
All rights reserved.
1

3. The 3rd of 3 Sessions
1. Intro and research design
2. Research instruments and data collection
3. Data analysis, interpretation, and presentation
2

4. 3
 Take care of yourself & ask for what you need
 Ask questions (some research lingo may be
unfamiliar and we are here to break it down)
 Step up, step back (draw out others’ ideas)
 Please try to stay off cell phone and social media
 Acknowledge everyone’s experience is different
 “One mic” (please mute phone when not
speaking)
 Session is being recorded
Agreements

5. Today’s Agenda
1. Analysis and Presentation
2. Interpretation and Conclusions
4

6. 5
1. Analysis and Presentation

7. 6
 Unlikely you will be running the analysis entirely by
yourself!
 But, you may have the opportunity to:
 Code unstructured data
 Have input into the types of analysis to run
 Review the analysis results, and help interpret the findings
 Suggest changes to the way the analysis is run
 To do those things, you need a basic understanding of
how to read and interpret tables, charts, and other types
of data analysis output
Data Analysis – the Patients’ Role

8. 7
Method
Selection &
Sample
Selection
Research
Question
Data
Collection
Data
Analysis &
Presentation
http://www.pcori.org/sites/default/files/PCORI-Methodology-
101-Training-Booklet-and-Resource-Guide.pdf
Research Steps

9. 8
•Level of education influences smoking behavior
•Weather has an impact on joint pain
Informed best guess
about what a study
may find
•Medicine A has better outcomes than medicine B
for chronic pain
•Medicine A has fewer side effects
You can have
multiple hypotheses
•Design your research study to test whether your
hypothesis is true. Do your observations support
your hypothesis or not?
•Test your hypothesis using statistical methods
•Draw conclusions
Test, Test, Test
Research Hypothesis

10. 9
• Examples: Rates (incidence, prevalence); close-ended
questions
• Discrete Data: data in whole numbers (e.g., 47 years old)
• Continuous Data: measures on a continuous scale (e.g.,
47.32 years old)
• Categorical: Descriptive data (no ordering, categories,
preferences)
• Ordinal: data that has some kind of ordering like age or
income level
Quantitative Data

11. 10
Qualitative data is based on narrative information,
not numerically ‘measurable’ information (e.g.,
“What does age 47 feel like?”)
•Perceptions
•Quotations
•Experience/Observations
•Open ended questions
Qualitative Data

12. 11
Health researchers want to understand the effect of a
medication on people with a certain condition
1. Complete a survey of their
symptoms before and after
taking the medicine.
2. Measure their symptoms before
and after taking the medicine.
1. Interview patients & ask them
to talk about their experience
when taking the medicine.
2. Conduct social media
discussion about patient
experiences
Source: http://www.pcori.org/sites/default/files/PCORI-Methodology-101-Training-Booklet-and-Resource-
Guide.pdfthe medicine.
PCORI Example
Qualitative methods
Quantitative methods

13. 12
Tools to Collect and Present Data
 Data can be presented in different formats that include
tables, graphs, narratives.
 Statistics is a tool that is used for quantitative research.
 Qualitative research uses non-statistical tools.
 Graphs can be used to present both qualitative and
quantitative data.

14. 13
58
60
62
64
66
68
70
72
74
0 50 100 150 200
Graphs: A Useful Tool to Communicate
H
e
i
g
h
t
Weight Source ://www.drugabuse.gov/publications/drugfacts/hivaids-
drug-abuse-intertwined-epidemics

15. 14
Data File Formats

16. 15
Quant Data File Example
ID Drugprice Drugtype RA OA Whypharm
1 $99 Generic Yes No I like the convenience
2 $99 Generic Yes Yes
Cheap, easy to get
there
3 $99 Generic No Yes Convenient, friendly
4 $99 Generic No Yes Open 24 hours
5 $199 Brand Yes No Near my house
6 $97 Generic No No Free delivery
Unique
identifier
Categoric
variable
(Single
Response)
Numeric
variable
Categoric variables
(Multiple Response)
Open-ended
Variable
• Columns represent questions (variables)
• First row shows question (variable) names
• Remaining rows represent individual “cases” (participants, respondents)

17. 16
Qualitative
• Transcripts (from
focus group
discussion or
interviews)
• Audio and video
recordings
• Social media posts
Data File Formats
Social Media Chat: Can Pets Help
People with Chronic Illness?
Getting my puppy has made so many
days so much better.
My dogs are my heating pads, my
laughter, my secret keepers, my
exercise, my bed warmers, my heart and
soul. They're there at night when I need
to cry over the pain without judgment.
My cats, horse, and pup give me a
reason to get out of bed on the tough
days.

18. 17
 Not computer programming “coding”
 Rather: Sifting through verbatim responses in patients’
own words, and “bucketizing” the results according to
themes you detect
 Used for both qualitative and quantitative research
 Patient insight can be very useful here
 But also could introduce bias, so be aware
Qual Data Analysis: What is Coding?

19. 18
ID Whypharm Convenient Price
1 I like the convenience 1
2 Cheap, easy to get there 1 1
3 Convenient, friendly 1
4 Open 24 hours 1
5 Near my house 1
6 Free delivery 1 1
7 Low prices 1
8 Can walk there 1
9 Coupons 1
10 In my grocery store 1
Coding - Example

20. 19
 What do you think of this website?
 The graphics are appealing
 The material is interesting
 I couldn’t find what I was looking for
 The font size is a bit too small
 I like the variety of information
 The information is presented in a way that is easy to
understand
 I like the overall look
 It’s hard to see on my smartphone
Discussion - Coding
What codes would you create for the following?

21. 20
Qualitative Analysis - Word Clouds
Source: BJC Health Connected Care
http://www.bjchealth.com.au/connected-care/2013/05/bjc-rheumatology-word-art/

22. 21
Basic Analysis Methods

23. 22
 To analyze
quantitative data we
use statistics
 Research question(s)
and data types dictate
analysis and statistical
methods
Data Analysis
N = Population size
n = Sample size
p = Statistical significance
(How likely differences
between groups are due to
chance)
p <0.05 means less than a 5% likelihood (probability) that study
findings are due to chance (and therefore, a greater than 95%
likelihood that the interventions are related to the results)

24. 23
 Statistical significance can be manipulated:
 http://fivethirtyeight.com/features/science-isnt-
broken/?ex_cid=538fb
 Statistically significant differences are not always
important research findings*
 Clinical significance tells whether research findings are of
practical use to patients and clinicians
Statistical Significance – A Couple of Caveats

25. 24
 Best for understanding “typical”, but still need to
understand distribution/range
 Use t-tests to compare results across groups
Measures of Central Tendency
The arithmetic average of a series of values
Mean
Median
Mode
The middle value in a series
The value that occurs most often in a series
Example:
2, 2, 2, 4, 6, 8, 9
4.7
4
2

26. 25
A way of summarizing data by describing how spread out
the data is.
The average score in a class on a math test may be 65
points. But we may also want to talk about the higher
scores (above 65) and the lower scores (below 65).
Measures of Spread can be calculated using different
statistical methods such as Standard Deviation and
variance.
Measures of Spread

27. 26
Frequency Data Tables
The number of observations
in each category is called
frequency.
Graphs
X-Axis = Categories
Y-Axis = Frequency
Data: Tables & Graphs
Weight in
grams
Frequency %
0-19 3 10%
20-39 14 15%
40-49 8 30%
50+ 12 45%
0
1
2
3
4
5
6
7
Blood Grp A Blood Grp B Blood Grp 0 Blood Grp AB
Blood Groups

28. 80
260
80
0
50
100
150
200
250
300
< $100 $100-$199 $200+
27
Distribution Example & Discussion
How would you describe the results of this study?
Drug Price ($)
Frequency
Price Paid for a One-Month Prescription
(n=420)

29. 80
100
0 0 0 0 0 0 0
60
100
40
20 20
0
20
40
60
80
100
120
<100 $100 110 120 130 140 150 160 170 180 190 200 210 220+
28
Central Tendency Example & Discussion
Mean = 153
Drug Price ($)
Price Paid for a One-Month Prescription
(n=420)
Now, how would you describe the results of this study?
Frequency

30. 29
Crosstab Example & Discussion
Change in Perception of Pain After 3 Months
Exercise Group
(n=300)
Control Group
(n=300)
Increased 10% 12%
No change 14% 66%
Decreased 76% 22%
Total 100% 100%
How would you describe the results of this study?

31. 30
2. Interpretation and Conclusions

32. 31
Describing the data is the first step
Interpretation involves relating results to the research
questions, and sometimes, to previous research
Ultimately, what do you recommend based on the
results?
•How might this affect the target population? (Patient perspective
can be key here)
•What other research is needed?
What Does It All Mean?

33. 32
There are a lot of things we didn’t cover, or just
touched on
• PCORI 101 Training Booklet covers a lot of good
material – some overlap with this training, and some
new
• We plan to continue the conversation…
Wrap-up

34. 33
Thank You!
By Ashashyou (Own work) [CC BY-SA 4.0
(http://creativecommons.org/licenses/by-sa/4.0)], via Wikimedia Commons

35. 34
 PCORI Methodology 101: Training booklet for patients and
stakeholders: http://www.pcori.org/sites/default/files/PCORI-
Methodology-101-Training-Booklet-and-Resource-Guide.pdf
 Gapminder.org: This resource has some very creative graphs and
resources to use statistics: http://www.gapminder.org/
 CDC: Data and statistics http://www.cdc.gov/datastatistics/
 Australian Bureau of statistics: A short video on statistical language:
http://www.abs.gov.au/websitedbs/a3121120.nsf/home/statistical+lan
guage+-+quantitative+and+qualitative+data
 Qualitative Health Research (QHR): This is a peer reviewed journal
on qualitative research: http://qhr.sagepub.com/
Useful Resources

36. 35
 Graphics inspiration
 http://www.washingtonmonthly.com/magazine/mayju
ne_2011/features/the_information_sage029137.php
‘Extra Credit’