Testing how many surveys are needed to get reliable and precise estimates from mobile phone surveys.

What is it about?

Assessing people's thoughts, feelings, and behaviors using brief smartphone surveys is an increasingly popular tool in behavioral research and clinical work. In clinical work in particular, the goal is to analyze the data that comes out of these surveys to get various statistics (mean, variability, correlations) that summarize an individual's behavior. For this to be a useful technique, these statistics need to be reasonably precise. To get precise statistics, sufficient numbers of observations (i.e., completed surveys) are needed. However, at this point little is known about how many observations. This study used simulations based on parameters drawn from real data to determine how many observations are needed. We found that for the mean and variability, most current study protocols would have sufficient observations, on average. But when it comes to correlations, many more observations than are typically recorded would be needed. The publication provides more detailed information about the necessary number of observations under several conditions. This should help researchers and practitioners collect the right number of observations for their assessment goals.

Featured Image

Why is it important?

The publication provides more detailed information about the necessary number of observations under several conditions. This should help researchers and practitioners collect the right number of observations for their assessment goals.