Data Collection 101
- Purpose - What do I want to learn?
- Survey Design - Framing the question: What should I collect?
- Protocol - How will I collect it?
- Processing - How will I process the information?
- Analysis - What insights can I gather? What does the data suggest?
- Distribution - How will I share the data?
Every data collection process is composed of several distinct steps. From the initial scoping and design to the final results, understanding the overview of your workflow helps you comprehend the entire start-to-finish process. With a successful project design, you should have a good idea of not only what you want to collect, but also what you expect to do with the data at the end.
Purpose
What do I want to learn?
Having a clear and defined purpose for your field data collection project is imperative. On-the-ground surveys can be expensive, both in terms of time and money. You should clearly define your goals and objectives before conducting your data collection activities.
Begin by defining your purpose by analyzing precisely what data you need to collect to answer your question. There will likely be additional data you could collect while in the field but take caution against scope creep2, or adding new work not contained in the original project. Any additional data collected can quickly increase costs and time to conducting and completing your survey.
If what you're looking to learn can be determined back in the office, don't include it in your field data collection process. For example, why collect information that can be pre-loaded into your survey for collectors to validate, rather than collecting it from scratch?
Survey Design
Framing the question: What should I collect?
How you set up your data collection survey will be critical to its success. The way you design how observations are recorded will have a tremendous impact on the way the data is collected and how its later processed and analyzed. An effective survey is one that is specific and focused. You don't want to go on a tangent away from your original purpose. "Nice to have" information can quickly take away from the precious time it takes for your team to conduct the survey; it should be scrutinized as to whether it really needs to be captured.
For your questions, provide appropriate answer choices to pick from. This includes framing your questions fairly and being inclusive while also keeping them brief and direct. Consider how the data will be processed and analyzed as you're designing your survey. For example: do you want to allow a question to be answered with an open-ended response, or provide a range of structured answers to choose from instead?
Other things to consider surrounding your survey design include whether you will need supporting reference information while collecting the data. If a field user needs to confirm that they're within a certain area while collecting the observation, how will they be able to verify that? Design and logistics go hand-in-hand. For example, will you be working in an area requiring access to cellular connectivity? Detailed tips and suggestions can be found in a later section on Survey Best Practices.
Protocol
How will I collect it?
Your protocol should include training the personnel who will be collecting your data. The level of understanding and competency your field staff possess will be reflected in the quality of your data. It will also be beneficial to have a long-term project timeline and schedule of daily activities.
Identify what personnel will be on your data collection team and determine if they will need to work together. Consider the logistics to safely and efficiently perform your survey. For example, your project may warrant a team of two - one driver and one person collecting data.
You may require certain conditions regarding weather and seasonality to be met before you can or when you cannot conduct your survey. As you define your procedure you will need to determine if you'll be collecting the data while driving in vehicles, on-foot, or perhaps even remotely. Establish a schedule if there will be any revisit frequency to your data collection.
Processing
How will I process the information?
After you collect a batch of data, it may require processing for your eventual analysis or visualization. This step in the process can vary greatly depending on how the data is being collected and what's required to make the raw data usable. Placing emphasis upon optimal design and an effective protocol step will benefit the processing piece of your workflow. If your protocol dictates that data will be collected with paper and pencil, you will require a post-collection processing task where results are transcribed into a computer spreadsheet or database.
Each additional step in the processing portion of data collection can introduce opportunities for error. Whether it's human error while transferring data from paper to electronic form, or using incorrect parameters in a programmatic conversion - each additional step added introduces an opportunity for error.
Analysis
What insights can I gather? What does the data suggest?
After defining the original purpose for collecting, gathering, and processing your data, you will need to analyze your results. Diving deep into the information collected from your survey provides the answer to what you were originally motivated to find out - and sometimes more! A thorough evaluation of the findings from your data may suggest conclusions which were only hypothesized before. If your analysis can replace subjective decision support with measurable facts, you will have more confidence in your conclusion.
Distribution
How will I share the data?
After your data has been processed and analyzed, you will likely want to distribute or publish the findings. There are various methods to convey your results: tabular, map, charts, graphics, etc. Often, a combination of methods will be helpful to tell your story but it is important to use the right approach in order to achieve maximum effectiveness.
Not all data collection workflows are created equally! Some tools are built upon platforms which allow your data to be instantly distributed in a variety of formats once it's collected. There are also other downstream integrations to consider, such as the ability to publish your data once and then connect it to other 3rd-party services. Learn more about some of the integrations you can do with Fulcrum, on our Integrations page
1. Water sampling https://www.flickr.com/photos/iaea_imagebank/10723081213/ ↩
2. https://en.wikipedia.org/wiki/Scope_creep ↩
3. Irish Summer https://www.flickr.com/photos/23629083@N03/6098106484/ ↩