Field surveys are mission-critical but they are prone to all kinds of errors. The data set can suffer from missing values, wrong responses and even falsified answers and researchers have to spend time “cleaning” the data before analyzing it. This process can be time consuming. Worse, it can distort the data set–even to the point of making it statistically useless.
At FINCA, we live this situation every day — we collect millions of data points from remote communities all around the world, measuring things like household consumption, job-creation, and customer satisfaction. With a small team and very limited resources, we don’t want to spend our time manually cleaning every piece of data we receive, and we don’t have the luxury of collecting (and then throwing away) badly-conducted surveys.
We use the best open-source data collection software (Open Data Kit), which digitizes the entire research process and allows us to install some basic data quality controls in the survey forms. We customized the platform to actively analyze the quality of the data as it is being collected. Using a variety of statistical controls, ValiData flags suspicious survey data in real time, allowing our staff to take corrective measures when the survey is still in the field.
ValiData enables our field staff to produce very high-quality data sets, while our statistical experts can spend their time actually generating meaningful insights instead of dealing with flawed field surveys.
ValiData performs a robust cross-check on individual survey responses, using standard statistical criteria and customized validation rules. These rules ensure that all of our survey responses are within expected parameters. ValiData even conducts advanced machine-learning functions which observe the behavior of the surveyors themselves, flagging anyone whose conduct is out-of-sync with his peers.
If you would like more information about ValiData or would like to explore potential collaborations with FINCA, we would love to hear from you.