digital data collection · 2017. 5. 24. · –offline collection => can be used even in remote...
TRANSCRIPT
-
DIGITAL DATA COLLECTION
Fanny Parenton
May 2017
-
INTRODUCTION
• Conducting a study is a very long process:
– Design
– Develop
– Collect in field
– Submit to analyst(s)
– Validate
– Analyze
– Report
• We will focus on data collection and address this issue in a comprehensive way
2
-
OUTLINE
I. Data collection software and application
II. Use of digital data collection instead of paper-based forms
III. Making it work
IV. How to choose a data collection software?
4
-
I. What is a data collection software/application?
• Data collection software is a tool with which you can:
• Build a computer-assisted questionnaire
• Submit and complete questionnaire on different electronic devices
(laptop, smartphone, tablet)
• Save and centralize all the data collected
• Upload and export data under different formats
• Present data on a map or with graphs and make rough analysis
5
-
I. What is a data collection software/application?
• Examples of data collection applications
– Many data collection applications are available
• Open Data Kit, Go Formz, iformbuilder…
• Some are commercials programs but a lot are free
– We will focus only on two common and free ones: Kobotoolbox (KTB) and Epicollect (EC)
• These 2 applications are very close (same general architecture, similar interface) but some specific differences can be noted
• By getting familiar with these two software, the aim is to provide you:
A general understanding of data collection software
A means of comparison to choose the most appropriate tools for your studies
6
-
OUTLINE
I. Data collection software and application
II. Use of digital data collection instead of paper-based forms
III. How should I use it?/How does it work?
IV. How to choose a data collection software?
7
-
II. Why should I use digital data collection rather than paper form?
• Speed:
– No need to transcribe data from papers to computer (data-entry)
– Less time needed for validation (see Accuracy)
– Real time data transfer, centralization and even rough analysis
• Accuracy
– Reduce typing or transcription errors:
• Ex: Avoid having different entries for a same answer by using drop down boxes or check boxes
Specie: Mutton
Specie: SSpecie: Sheep
Specie: Ewe
8
-
II. Why should I use digital data collection rather than paper form?
• Accuracy
– Automatic and real time data validation
• By using input masks
Date : 2nd of May
Date : 17/05/02
Date : 02/05/2017
9
-
• Accuracy
– Automatic and real time data validation
• By using validation criterions
Question: ‘How many cows have been affected in your farm?’
As validation criterion we can put that the answer must be a number inferior to the total number of cows in the farm
EXAMPLE
II. Why should I use digital data collection rather than paper form?
10
-
II. Why should I use digital data collection rather than paper form?
• Data security:
– Data storage in a unique central database
– Automatic saving (no risk of loosing data between collection and transcription)
– Restricted and secure access to data
• Availability and sharability of data
– Rights to data access can be defined specifically
– Better communication between actors and between actors and stakeholder (real time availability of the data, possibility of knowing how work is progressing)
11
-
II. Why should I use digital data collection rather than paper form?
• Adapted for field work
– Offline collection => Can be used even in remote places
– Requires tools (tablet, smartphone…) most of the people are already familiar with.
• Cost:
– An investment cost at the beginning (to buy or make use of personal electronically devices)
– Less operational costs (no need to transcribe data from paper to computer, less time needed to clean up and centralize all data…)
12
-
II. Why should I use digital data collection rather than paper form?
A lot of advantages…
….however a few restrictions on the use
Digital data collection isn’t well adapted to some specific kind of study - Semi-structured interviews
- Non directive interviews
- Any interviews with many open ended questions or collective interviews
13
-
II. Why should I use digital data collection rather than paper form?
Speed
Accuracy
Data security
Data availability
Operational costs
Initial investment
costsNot adapted to non directive
interviews
FOR AGAINST
Balance sheet
14
-
LET’S TAKE A BREAK
15
-
OUTLINE
I. Data collection software and application
II. Use of digital data collection instead of paper-based forms
III. Making it work
IV. How to choose a data collection software?
16
-
III. HOW SHOULD I USE IT? HOW DOES IT WORK?
A. Zoom in on some important functionalities1. Form builder interface in EC and KTB
2. Questions settinga. Different kind of questions
b. Different ways of assembling questions
3. Assigning rights to access
B. Main steps to conduct data collection with software
17
-
Type of question To access question
setting menu
To access group
setting menuName of the group
New question
III.A.1. Form builder interface
18
-
Video 1KTB1: Kobotoolbox form builder interface
19
-
III.A.1. Form builder interface
All the different types
of question available
Setting of the selected questionQuestionnaire
Type of question
20
-
Video 2EC1: Epicollect from builder interface
21
-
III. HOW DOES IT WORK?
A. Zoom in on some important functionalities1. Form builder interface in EC and KTB
2. Questions settinga. Different kind of questions
b. Different ways of assembling questions
3. Assigning rights to access
B. Main steps to conduct data collection with software
22
-
III.A.2. Questions settings
a. Different kinds of questions available:
In all data collection software, we find the same major types of questions:
– Text
“Owner name?”
– Numeric (Integer or decimal)
• Decimal: “Milk production?” eg. 18.7kgmilk per day
• Integer: “Number of cows?” eg. 34 heads of cattle
– Date & time (With automatic input masks)
23
-
– Simple choice question
“Species of the animal sampled?”
– Multiple choice question
“Species present in the farm?”
– GPS (Geographical position)
• Register automatically the coordinates
– Photo and video
• You can add photo of clinical lesions, register, bill…
24
-
Choice menu in EC and KTB
2525
-
Video 3KTB2: How to create a question on Kobotoolbox?
26
-
Video 4 EC2: How to create a question on Epicollect?
27
-
b) Different ways of assembling/setting questions
– Conditional setting
- Some questions may not apply in every interview. (You may have a long list with a lot of questions and just some of them have to be answered)
Tedious work, potential error while filling and transcribing the questionnaire
- With data collection software you have the possibility to have only these questions appear that are relevant to the particular situation.
28
-
Symptom Y/N Nb of cow
Nb of sheep
Nb of goat
Starting date Ending date Treatment
Lameness
Salivation
Mouth lesion 12 3 12/02/17 01/03/17
Foot lesion
Loss of appetite
Milk drop 2 15/02/17 15/03/17 Yes…
Etc…
We are interested by the FMD symptoms observed. Depending of the symptoms we would like to ask for general and specific details
With a paper form:
EXAMPLE
Specific details missingMilk production before, average milk price Etc…
Many empty boxes
A large table with many columns and rows
Confusion Potential errors
29
-
We are interested by the FMD symptoms observed. Depending of the symptoms we would like to ask for general and specific details
With a Kobotoolbox: First, make a check box question with the list of all the symptoms Then, make a list of conditional questions for each symptom
EXAMPLE
You can adapt for each symptom the questions list
3030
-
III.A.2.b The different ways of assembling questions
• Conditional setting
– In Kobotoolbox:
• You can set for each question a criterion (based on a previous answer) to make this question appear or not (as previous example)
– In Epicollect:
• You can add ‘jumps’ from one question to one other (you skip others questions)
Q1) Species present in the farm? Ovine Bovine Caprine OtherQ2) If other, please specify:………………..Q3) Last vaccination against FMD?
In Epicollect: If the answer isn’t “other” jump directly to Q3(on KTB: Condition for Q2 would be ‘other’ ticked in Q1)
EXAMPLE
31
-
• Question grouping
• Several questions can be gathered into a group.
• Some specific setting can be applied to the whole group or to only one question of the group.
• Technical difference while building the form:– On Kobotoolbox: You can move a question into or out of a group
– On epicollect: You create questions in the group directly and can not move them after
32
-
– Repeat setting
• One question (or one group of question) can be repeated as much as needed
• Available with Kobotoolbox
EXAMPLE
We are interested by animal which died of FMD, we would like to know for each one: Species, Age, Value of animal, Date of death
We can put all this questions in a group and add the repeat option for the whole group
33
-
Question settings
34
-
III. HOW DOES IT WORK?
A. Zoom in on some important functionalities1. Form builder interface in EC and KTB
2. Questions settinga. Different kind of questions
b. Different ways of assembling questions
3. Assigning rights to access
B. Main steps to conduct data collection with software
35
-
III.A.2 Rights assignations
• In data collection software you can choose:– Who can modify the form
– Who can view the data collected
– Who can collect data
• In Epicollect
- 3 user status differentManager: Can setup/modify/amend project
Curator: Can view and collect data
Collector: Can only collect data
36
-
• In Kobotoolbox
Rights assignations
37
-
TIME FOR ANOTHER BREAK
See our next presentation about:
Main steps to conduct data collection using software
How to choose the software that suits my purposes?
38
-
III. HOW DOES IT WORK?
A. Zoom in on some important functionalities1. Form builder interface in EC and KTB
2. Questions settinga. Different kind of questions
b. Different ways of assembling questions
3. Assigning rights to access
B. Main steps to conduct data collection with software
39
-
Main steps to conduct a data collection with software
1 Design survey
2 Develop questionnaire with software
Build digital form
Try it out
Assign rights
3 Collect & centralize the data
Train field interviewers (enumerators)
Conduct questionnaire
4 Upload, validate and analyze data
40
-
Step 1: Design the survey
Very important step!!! Do not do it too quickly!
- Write a draft paper :
Define your objective first
Second question: what data are needed to answer the study objective?
Taking care to:
Target the data: Do not collect useless data
Prioritize the data: Focus on vital data but also collect data that could be interesting for the analysis
Cross-check the data: to avoid error
More details on this step and on the pitfalls to avoid in a upcoming training
Main steps to conduct a data collection with software
41
-
Step 2: Develop the questionnaire on software by three steps
- Build digital form
- Try it out
- Assign rights
Main steps to conduct a data collection with software
42
-
Step 2: Develop the questionnaire on the software
- Build digital form
- Try it out
- Assign rights
Main steps to conduct a data collection with software
I have my draft paper questionnaireHow do I ‘translate’ it into a digital one?
Let 'see one example
43
-
Step 1: What kind of question is it?
4444
-
Video 5KTB3: Show the questionnaire
45
-
Step 2: How are questions grouped? The logic pattern(Example with KTB)
4646
-
One way of doing it with KTB:
- We have 8 questions repeated 3 times for different species:
- 1 multiple choice question
- 7 numeric (decimal or integer)
-Step 1: Copy questions and make 3 groups for different species
-‘Impact on bovine herd’-‘Impact on ovine herd’-‘Impact on caprine herd’
-Step 2: Create 2 sub groups-‘Dead animal’-‘Milk production’
-Step 3: Set the conditional setting of the groups and sub groups
-> Group ‘impact of ovine’ must appear only if ‘ovine’ was checked at the question ‘herds concerned?’
-> Subgroup ‘dead animal’ must appear only if ‘Death’ was checked at the question ‘Symptoms observed?’
47
-
Videos: - 6KTB4 How to copy question?- 7KTB5 How to create group?- 8KTB6 How to set the group option?
48
-
Step 2: How the questions are grouped? The logic pattern
A group consisting of 5 questions.The whole group can be repeated as much as needed
4949
-
Video 9KTB7: Veterinary expenses question
50
-
Step 2: Develop the questionnaire on the software
- Build the digital form
- Try it out
- Assign rights
B. The main steps to conduct a data collection with software
Mandatory!!To be sure that everything is functioning well(Take into account all the scenarios possible)
51
-
Step 2: Develop the questionnaire on the software
- Build the digital form
- Try it out
- Assign rights
B. The main steps to conduct a data collection with software
You have to choose who can:- Collect data- View the data (Stakeholders?)- Modify the project
52
-
Video 10KTB8: Rights assignmentVideo 11EC3: Rights assignment
53
-
Main steps to conduct a data collection with software
1 Design survey
2 Develop questionnaire on software
Build digital form
Try it out
Assign rights
3 Collect & centralize the data
Train field interviewers (enumerators)
Conduct questionnaire
4 Upload, validate and analyze data
54
-
B. The main steps to conduct a data collection with software
1 Design survey
2 Develop questionnaire on software
Build digital form
Try it out
Assign rights
3 Collect & centralize the data
Train field interviewers (enumerators)
Conduct questionnaire
4 Upload, validate and analyze the data
55
-
Data format export
Epicollect - Download under CSV or JSNO
- Show data on a map
Kobotoolbox
-Download data under XLS, CSV, ZIP...
- Show data on a map
-Show graphs
56
-
Data in XLS format
If there is a group/question repeating -> Additional tabs
57
-
Video 12 EC4: How to present and upload data on Epicollect?Video 13KTB9: How to present and upload data on Kobotoolbox?
58
-
OUTLINE
I. What is a data collection software/application?
II. Why should I use digital data collection rather than paper form?
III. How should I use it?/How does it work?
IV. How to choose a data collection software?
1. Summary table of the comparison between KTB and EC
2. Points of comparison that you can use to choose the software the most appropriate to my project
59
-
Epicollect Kobotoolbox
Data export format CSV or JSNO(Excel compatible)
CSV, XLS…(Excel compatible)
Data collect possible even within internet access? (offlinemode)
Yes
Right designation/ access setting functionalities Almost the same
Type of questions/specific functionalities
Major type of questions The same major type of questions (text, numeric, GPS, photo, date…)
Drop down box:Possibility of entering a list of answersin Excel/CVS format without having tocopy one by one all the possibilities?
No Yes(useful for location names)
Input mask for text question We can choose only between: only numeric / only letters
More detailed: We can choose to have x letters then y numbers (could
be useful for example for the id of animals or blood samples)
Validation criterions Can only compare to a fixed value Can take into account a previous answer
Assembling question functionalities -‘Jump’- Create ‘branch question’
- Occurrence criterion- Repeating question
More flexible
Summary table of the comparison between KTB and EC
60
-
• Points of comparison that you can use to choose the software the most appropriate to my project?
– Who is going to collect data? With which digital device?
– In which place the data will be collected? (Offline mode needed?)
– Are there very specific functionalities that can be useful? (depending from the data you want to collect)
– Do I need to have a continuous monitoring of the data collected?
– What kind of export format I need?
– …. Not exhaustive list depending from your project!
61
-
Conclusion
Data collection application are good tools that, when adopted well, can help you collect good quality data and save a lot of time
But it is important to be careful and take time to choose the application that is most convenient for your study and to develop the data collection
The first time you use digital data collection, you may want to use an old-fashioned paper copy as well
THANKS FOR YOUR ATTENTION62
-
Links
• Kobotoolbox website http://www.kobotoolbox.org/
• Epicollect website http://www.epicollect.net/
• E-learning on data collection and kobotoolbox with free access (French or English)
https://elearning.cirad.fr/course/index.php?categoryid=23&lang=en
http://www.kobotoolbox.org/http://www.epicollect.net/https://elearning.cirad.fr/course/index.php?categoryid=23&lang=en