KoBo—Mobile Data Collection in the Field

Post image for KoBo—Mobile Data Collection in the Field

by Maryann on December 9, 2012

This post is an introduction to the KoBo Toolbox, a suite of software programs which allows the user to create and complete survey and other data collection using cellular phones and mobile (Android) devices.  I used  KoBo for my research, and my goal for this post is to talk about why anthropologists should use it, how it can help them, and how they can get started.  I also give field-level experience and explanations.  I’ve actually used the software for two different research projects—for a quick review of how I felt about the experience after my first project see my blog post for ODK from 2011.  That one focuses mainly on managing a team of researchers and how KoBo helped with that.  For this post I’m focusing on my last project (2012) and how KoBo is useful for anthropological field work.

This post is divided into 4 different sections: 1) What is KoBo?  2) Why should anthropologists use KoBo? 3) How did KoBo work in practice? and 4) What does a researcher need to get started with KoBo?   I would also like to highlight that KoBo is both opensource and FREE.  That means there is a community of researchers and programmers ready to help you out and you don’t have to pay for the software.

Section 1) What is KoBo?

KoBo is a set of software that allows you to complete research using mobile, android devices.  KoBo is an implementation of ODK collect, a base code for survey programming.  This is an example of open-source, non-competitive cooperation, and the result is just awesome.  KoBo toolbox has implemented several supporting pieces of software that aid the researcher in creating assessments, collecting data, syncing that data to a format from which it can be analyzed, and mapping that data, too.  KoBo supports a variety of information types, including text, numerical, single answer, multiple answer, cascading questions, geopoint, audio, video, photo, and barcode data.  On android cell phones.  No paper.  KoBo is supported by the Harvard Humanitarian initiative.  It has been field-tested and researcher-approved.

So, here are the goods:

  • KoBoCollect is a piece of software that allows you to collect data using mobile phones.
  • KoBoForm allows you to create, edit, and export forms for data collection and assessments (e.g. surveys).
  • KoBoSync allows you to synchronize and aggregate your data to HTML and CSV formats (e.g. you can read your data in an excel spreadsheet).
  • KoBoMap provides software that utilizes google maps so that you can show your data spatially.

I used the first three of these components while in the field (see section 3), and intend to use KoBoMap to present my data (stay tuned for a future post on that one).

Section 2)  Why should Anthropologists use KoBo?

As I see it, KoBo holds a wealth of benefits for anthropologists.  It has the potential to assist researchers in interviewing, focus groups, community mapping, photo essay/photo elicitation, video/participatory video, and rapid ethnographic assessments.  The researcher creates their own format based on their own needs—so, if you need a combination of Likert scale questions along with a couple of in-depth interview questions, a couple of photos, and a GPS point, you can do that.  All that demographic data you need for each research participant?  KoBo can help you with that. Just need a map with photos and video?  KoBo can do that, too.

There are also field-based reasons why KoBo is a great choice.   KoBo was actually designed to be used after disasters and for human rights projects.  KoBo is a good choice for disaster research when you need to get data gathered, synced, and analyzed quickly.   It is also appropriate for when a full-scale paper project may not be feasible for reasons of confidentiality.  You can gather and send your data, leaving no paper trace of the responses in case of sensitive material.  You can get data out of a country where protection of participants may be extremely important (without having to explain yourself at the airport).  If you are working with non-literate populations, KoBo supports non-text questions.  You can have the question be voice or a picture, and then you can have the answer be recorded in voice or a picture, too.  Even though, yes, it is a cell phone—I think it is a great tool for lessening the technological divide between researcher and respondent.  Everyone has seen a cell phone, so it is a non-flashy way to gather and record data.

KoBo is also excellent for research teams.  KoBo allows for question and answer formats to be constrained or standardized.  So, say you only have IRB clearance for respondents over 18—if your research assistant enters an age of 17, you can have the software pop up a warning that the researcher should not continue (and actually bar them from doing so).  There are data constraints and logic for numerical questions (so you can’t be 1,000 years old, for instance—by setting the range from 0-107).  Furthermore, there are separate interfaces for staff and supervisors, which limits permissions to change, delete, or send finished data. Staff member can’t make changes to forms or completed data.  In my experience I had better accuracy and a lower error rate because of these data maintenance features.

A researcher can also change their survey or assessment mid-project.  Need to change a translation of a word to the local dialect of Spanish (instead of the kind that you learned, originally)?  You can do that, and have it changed for the next version of the survey within 15 minutes.  The program supports multiple languages (so I had the same questions in an English and a Spanish option for my research).  You can check your data daily, looking for errors or mistakes (I had one research assistant that seemed to always press the wrong town location for at least one survey a day)—but I knew where they were that day and could catch the mistake instantly.  Finally, there is near-instant aggregation of your survey data into a single file.

Want more benefits?  Once your data is collected, you can instantly (seriously, 15 seconds) aggregate your data into an excel spreadsheet/CSV file.  You can also get an instant (LABELED!)  SPSS file with an SPSS labeling function (part of KoBoForm).  And mapping is a breeze—you can collect GPS points and look at your data/responses topographically.  Excited yet?  KoboMap allows you to create maps with data display layers and a customizable legend. For example, your map might show different counties shaded brighter or darker according to the reported level of access to water.

Section 3) How did KoBo work in practice?

I used KoBo, exclusively, for my surveys in the field.  I have worked for the past two years in Alto Beni, Bolivia on research related to water and sanitation (with a specific focus on wastewater and wastewater treatment technologies).  Both times I used KoBo, and both times it worked well.  I will share a bit about how I used the software, how it worked, and what it was like to implement research using KoBo.

I mainly used KoBo for a structured survey and for GPS mapping.  I programmed the survey using KoBoForm, used KoBoCollect to gather data on Android phones, and used KoBoSync to get my data (nicely labeled!) into SPSS.  I’m still analyzing the data now—but I intend to use KoBoMap to present my data.  Each of my surveys was about 100 questions long and included a combination of text entry, numerical, multi-select, and single-select questions.  I also took GPS points at each location so that I could map answers and link them back to topography/neighborhoods.  I had 5 android cellular phones (Huawaei Ideos model—approximately 130$ each), and several (awesome) research assistants.

Prior to the Field:

First, I had to create the KoBo form.  Best practice for this is to create your survey in Excel using a template.  I wrote and edited the survey in Excel.  After this, you’re ready to start with KoBoForm.  During my first trip to the field there was no KoBoForm and I coded the survey by hand.  That took a few days.  Now, with the form builder, depending on the size of your survey, you could finish in a matter of hours.  For me, since my survey was lengthy, it took a couple of days (but this is the kind of stuff that you can do watching Battlestar Galactica reruns, so no worries).  For web tutorials on this part of the process see the KoBo website.  Then, I installed KoBoCollect onto the phones, and loaded the forms.  Always go into the field with working, completed forms, if possible.  You could do it in the field, there is an offline version of KoBoForm, but it is easiest if you do this in advance and save your precious field time for data collection.

In the Field:

Using KoBo in the field was great overall.  However, it was a process that required planning and forethought.  And a ton of hardware! Getting through customs with 5 cell phones, a variety of wires, two laptops, and a bunch of other related tech requires a bit of an explanation.  If you are completing research using KoBo somewhere without regular access to electricity, you’ll need a solar charger.  Luckily, we had regular access, so I only brought a couple in case of outages.

KoBo also requires training—for you and for others if you are going to employ research assistants.  You will need to make sure that you have knowledge of all of the KoBo components, know how to trouble-shoot, and can do a bit of rudimentary XML forms (X-forms) programming if you get into a pickle.  You can get all of this information and training from the KoBo website.  Depending on the technological savvy of your assistants, training them just to use this software on the phones could take anywhere from one day to one week.  The research assistants in my program required about 1 day of training.  During training you go through all of the steps with loading the form, asking questions, entering correct answers, and skip logic.  Obviously, if you are an anthropologist, there is other training to be done related to human subjects, informed consent, etc. –but for KoBo only, leave a day.  Assistants need to practice taking the surveys, saving them, etc.  If you are completing research by yourself, you can just hit the ground running because you’ll have studied up before the field.

Naturally, you may have to make changes to your form after an initial pilot.  This is pretty easy.  Is there a specific term used in your site that isn’t standard Spanish, for instance?  You can just go back, change it in KoBoForm, and reload it.  Do you need to add a neighborhood in a multi-select answer?  No problem.  My fixes took a couple of hours to tweak and reload.  All in all, training, pilot, and prep required 2 days of work upon arrival to the field.

Once the initial bugs were worked out, the data collection phase went smoothly.  All the research assistants loved the technology, it is fast, phones are light, and the tech is low-profile.  Researchers asked the questions and then entered the individual responses (for my research, oral surveys were much more appropriate than written).  All of the RAs had my cell number in case of any issues—but with good training there wasn’t much of a learning curve.  You don’t have to carry any paper with you do to do this research, so hiking up mountains, trekking through mud, getting chased by wild boars, and other such fun wasn’t hindered by a satchel full of paper surveys.  I got everyone involved in the research a phone case, and that was a great idea. Respondents barely even commented on the technology, and if they did it was usually because they thought it was cool. It was a great, fast way to collect quite a bit of data.  The GPS is a big selling point for me, too.

At the end of the day, I would collect all of the phones, sync (using KoBoSync) all of the answers to my computer, and check for errors or issues.  You don’t have to look through individual surveys because all the data is in one file you can review in Excel or SPSS, so it’s not hard to find errors like the one research assistant who entered the name of the town wrong once a day. I would also do quick look-throughs of the data to determine preliminary trends and stay abreast of the responses.  This did take some time—syncing the 5 phones each night took about 25-30 minutes, and then I had a rather extensive back-up routine for the outputs.  Expect to spend at least 45 minutes each day on maintenance for your data backups and syncs.  For me, it was a good way to decompress and then get ready to write some field notes.  I would say the benefits of being able to see your data each day outweigh this extra time, especially when you take into account that you don’t have to do any data entry later!

There are some things to think about when completing research with KoBo.  Security is a bit of a big deal with such a large amount of technology.  I had a lock box where I kept each of the phones when I was out of the room.  Charging is also a big deal, you’ll want to make sure that you have adapters and back-up adapters and a step-down converter if the country you are in has a different voltage.  The best advice I got before entering the field was this: “Think of the things that, if you didn’t have them, would ruin your research.  Bring two of those things.”  (Thanks, Micah!).  So, don’t go with one phone, one power cord, or one laptop.  Bring a couple of those things.  You need them.  Admittedly, if you didn’t want to sync your data until you got home you could do that. But lose a phone, you lose your data!  Of course, my computer crashed the DAY before I went to the field from La Paz.  I ended up getting it to work, but I was glad to have a backup netbook just in case.

I didn’t use the video, photo, or voice functions in my research in Bolivia.  However, that was more due to the fact that I had an awesome, hi-res GPS camera for photo and no need for voice or video in my research.  These functions work and work well; I have tested them through other projects.

 

After the field:

Admittedly, I’m still in the data analysis process.  But my survey data was in a beautiful, labeled, SPSS database before I even got back from the field (including GPS points).  I’ll be using KoBoMap to do some mapping.  However, for anthropologists it is important to ensure that using the GPS points doesn’t directly identify your respondents.  So, I’ll be using sections of data rather than household-level data for presenting results.  If you keep data on your SD cards from the phones, be sure to remember to keep them locked up with your other research components.

 

Section 4)  What do you need to get started with KoBo?

There is a bit of a learning curve with KoBo.  But it is nothing compared with entering all of your survey data by hand into an excel spreadsheet, so no worries.  For all the things I’ve discussed above and more, it is a great idea to learn and implement KoBo-based research.  You’ll need to really look into the KoBo website (kobotoolbox.org), and download KoBoForm, KoBoCollect, and KoBoSync, before you get started.  You can work with KoBoMap afterward if it is relevant to you.

Also, you will need to purchase some hardware.  Really, any android device will do.  But here are my recommendations:

  1. An android operating system.
  2. An unlocked platform (as opposed to a locked phone that only works with U.S. carriers, for instance).  This really only applies if you are doing international research and intend to also connect to a cell network.
  3. A pretty good camera if you intend to take photos (go for 3 megapixels or above).
  4. GPS.
  5. A phone, NOT a tablet.  First, these are still pretty buggy.  Also, they will (in my opinion) not be as low-profile or easily accepted by respondents.
  6. An SD card if your phone doesn’t have internal memory (a small one is fine (2 gigs) should be fine.  This stuff takes very little space.  If you are doing video, photo, or audio maybe kick it up to 8-16 gigs).

Optional:

  1.  A fingerprint reader.  This could be really cool for bioanth data collection, but of course you’ll have to consider the ethical issues and IRB application.

Accessories (*=depending on location) and other hardware:

  1. A working, reliable laptop.
  2. *A backup laptop or netbook.
  3. Extra power cords.
  4. *Step-down converter.
  5. *Solar batteries or chargers (Solios are pretty great)
  6. Adapters.
  7. *Lockbox.
  8. The email address of KoBo technical assistance:  info@kobotoolbox.org

I will probably never complete a paper survey again.  KoBo works well, is a great help to the researcher and research assistants, and offers a wide range of capabilities.  I highly recommend the software and would be happy to help anyone who is interested in using KoBo in the field.  There are definitely things to consider before using KoBo, but with a bit of work and preparation it is a streamlined and beneficial field tool for anthropologists.  I do want to note that I didn’t do any of this without help. The team at KoBo helped me out quite a bit with getting started on the software, gave recommendations on hardware, and was in touch with me while I was in the field.  A big thank you to the KoBo team!  And of course, a big thank you to everyone who worked with me in the field!

My email address is mcairns@mail.usf.edu.  Seriously, feel free to email me if you have questions.

“This material is based upon work supported by the National Science Foundation under Grant No. 0908425.  Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.”

 

From the Author: Hi, my name is Maryann Cairns.  I am currently a doctoral candidate in applied anthropology at University of South Florida.  Alan invited me to submit a guest post on the software that I used for my field research—KoBo.

{ 2 comments… read them below or add one }

Alan Perry December 12, 2012 at 11:10 am

Awesome work! But I’ve been assuming this was Alan Schultz. How about introducing the author at the start? I’m not sure I even know the author. You’re 95% of the way to awesome stardom. Anthrohacker is very, very cool. Check out Socialbrite.org for a good way to have guest bloggers. Keep up the good work.

Reply

ralleenwillems May 22, 2014 at 5:05 pm

Just recommended Kobo to some colleagues based on your article. Thanks for sharing the experience!

Reply

Leave a Comment

Previous post:

Next post: