Part three in our series on getting starting with mobile data collection focuses on the importance of defining and describing the data you need to collect.
Who wants to pay for features they don’t need?
The fact that you are even reading this means you are doing your research to avoid this common pitfall when selecting a mobile data collection tool.
However, with the number of features mobile data collection apps tout, it is often difficult to discern what you actually need and what your team can live without.
So how can you get rid of all the noise and focus on what you need? The key to selecting a platform that serves your needs precisely is understanding your data requirements. A clear understanding of your data requirements will help you narrow your focus and identify the right mobile data collection software for your project.
The first step in this process is to make a list of what data you know you need. Sometimes, you will have existing documents to give you a head start, such as results frameworks, M&E frameworks, or requests from supervisors and funders. Other times, you will need to start from scratch. In either case you should outline an initial list of data requirements, and then stress test that list to identify other factors. There are a number of questions you can ask that will help expand the list and describe each variable in more depth.
This post will explain how to:
- Categorize your data to understand what part of your project it influences
- Describe your data to uncover its attributes and characteristics
- Account for environmental factors to ensure you know how to collect the data you need
Health workers in India review the data they need to collect for their mLabour program.
How to categorize your data
Most often in mobile data collection and service delivery programs, your data can be broken down into two categories: (1) program performance metrics and (2) worker performance metrics. These categories help explain which aspect of your program the data affects.
Program performance metrics focus on how well you are meeting the project objectives. On the other hand, worker performance metrics are the best indicators for how well your workers are performing their duties and how much they are contributing to the success of the project.
When you want to know whether you are accomplishing your project objectives, examine your program performance metrics. How many beneficiaries have you reached? What percentage of your beneficiaries have improved health outcomes or crop yield? The answers to these questions are made up of many different variables, such as patient weight, disease contraction rate, or pounds of crops harvested, which will help you determine any improvements to beneficiary outcomes as a result of your intervention.
When you want to know how efficiently your team is working, take a look at your worker performance data. How long does it take for a data collector to submit their data after a field visit? How long does your team spend on data entry? How many beneficiaries does each worker reach and when? Keeping track of metrics such as number of house visits and form submission times will help you optimize individual and team performance.
How to describe your data
Once you have organized a list of your data requirements by category, flesh out their attributes and characteristics. There are numerous questions you can ask to help with this:
- Are you searching for longitudinal data–that is to say, are you looking to update the same metrics from the same source over time? This type of data requires a feature called case management (or long hours spent on data entry to collate your results).
- Does your data require outside data sources? Many governments have regular reporting on health, income, agriculture, and many other sectors. This is helpful when you are trying to compare your data to national averages, for example.
- Does one variable depend on another? For instance, before asking details about a patient’s treatment history, make sure that patient has actually received treatment. When you ask a patient if she has ever received medical treatment, and she replies, “no,” you can use skip logic (also known as display conditions) to avoid asking about vaccinations, medication, or other medical treatment.
There are many more questions you can ask to help describe the characteristics of your variables, but as with everything else, they will depend on your project’s objectives. Marcos Lavandera, health analyst at Pro Mujer, a woman’s development organization in Latin America, explained that for his project that focused on women’s healthcare in Mexico, his entire team took part in the process of defining their data’s characteristics.
“We had our program director, health analyst, and medical director all working together to make sure we looked at the data from all the possible perspectives,” Lavandera said.
All of the characteristics you define will help you later, as you determine the right tool for your needs and structure your mobile data collection forms.
A field worker in South Africa collects his latest round of data from a newborn and his mother.
How to account for environmental factors
Once you have listed, organized, and described all the variables you need to collect, you still need to account for where you are collecting data and who you are collecting it from. Pro Mujer recommends beginning this effort with the beneficiaries. By putting them first, you make sure they are the ones experiencing the greatest impact. Understand the data they can provide and the environment they live in to best provide the services you hope to offer.
Here are a few questions to consider when examining your project’s environmental factors:
- What are the languages spoken by the people involved (both data collectors and beneficiaries)?
- What is the reading level of your typical field worker or beneficiary?
- What is the level of mobile connectivity in the region? As a backup, is there a reliable WiFi connection when workers return home from a region with low or no mobile connectivity?
- How familiar are your data collectors with mobile devices? What is their level of digital literacy?
When you think the answers to these questions might pose a risk to your program’s success, it helps to speak with the workers on the ground. For instance, we have found that when we cannot get a phone signal, local mobile phone users often have a specific spot that works. Dimagi Senior Field Manager Nick Nestle encountered this issue on a field visit in Zambia:
“When we arrived at the village we realized we had no service on any of the carriers. This was a major problem, as the village was remote and travelling to get a connection would be very difficult. Our program hinged on the data being synchronized every day.
We started brainstorming ideas – could a more expensive phone get a signal; perhaps we could convince the mobile operators to expand coverage (not likely); should we buy the workers bikes to ride to where reception was? We didn’t have any real viable ideas when one of the local village workers overheard and stopped us.
‘No, no, no,’ she said. ‘Do you see that ant hill in the distance? That is where we get our reception. Every day, at five o’clock, I will go stand on that ant hill and hold my phone up in the air to synchronize. It will be fine.’ This was apparently a well-known solution in that village and one that everyone was used to doing.”
We recommend collaborating with the local workers to overcome challenges like these. You might know the right questions to ask, but they are the ones with the best understanding of the environment you will be working in.
Why does it matter?
With the myriad features and tools available for mobile data collection, developing a clear, written summary of all the variables you need will help keep you focused. Differentiating between program performance metrics and worker performance metrics will help keep your data organized. The characteristics of your variables will help determine how they interact in your forms, as well as how they are stored. The environmental factors associated with the region and people you are working will help determine the method and features of data collection that will work best for your program. It won’t surprise you to know that data is the most important piece of any mobile data collection program, so a comprehensive understanding of the data you need to collect is vital.
Once you understand how your current data collection process works and what data you need to collect, the next step in our starter’s guide to mobile data collection is learning how to determine the right mobile data collection platform for your program: “How to Choose the Right Mobile Data Collection Tool”