How are the Certainty and Margin of Error calculated?
When you hover over a datapoint in the dashboard, you'll see the following pop-up, which includes information on the Certainty and Margin of Error for that datapoint.
The traffic light system (High = green, Medium = yellow, Low = red) is intended to help you understand the quality of the datapoint at a glance, based on the Margins of Error.
The thresholds are defined as follows:
High: Margin of Error is less than or equal to 3%
Medium: Margin of Error is greater than 3% and less than or equal to 10%
Low: Margin of Error is greater than 10%
Margin of Error
We use MRP (learn more here) to generate the data you see in the Latana dashboard. This means the Margin of Error is calculated differently than when you look at a dataset of quota-sampled survey responses.
Rather than using a formula to calculate the Margin of Error, in MRP we generate a large number of estimates for every KPI in each segment (for example, aided awareness among urban, highly educated, medium-income, 18-25 year olds). The datapoint we surface in the dashboard is the mean of those estimates, and the margin of error is calculated using the lowest and highest of those estimates.
The meaning of the Margin of Error in MRP is the same as in other methodologies: it's intended to help you understand how good a datapoint is.
With MRP, the margins of error for the data we generate are smaller than those you would normally get with quota-sampled survey data. However, the margins of error still correspond inversely to the underlying sample size, so a small niche segment will have a larger margin of error.