What is data confidence?

Confidence surrounding the report card grades is measured on a five point scale. This tells us how confident we are, from low to high, that the calculated grade reflects the true condition of the indicator. Confidence is affected by many components, including the quantity and quality of data used in grade calculations.

We provide a confidence score for all of our indicators in each of our reporting zones. As indicator scores are rolled-up into final scores, so are the scores for confidence. For a detailed look at how confident we are of the grades for each indicator, visit our interactive results page.

What contributes to data confidence?
When reporting at a regional scale, one of the most important factors in determining how confident we are of a grade in a reporting zone is by considering how ‘representative’ the data is of the rest of the reporting zone. For example, when calculating a grade for water quality, we have higher confidence in a grade if data comes from multiple sites within a reporting zone, compared to a grade if data comes from only one site. Likewise, we have higher confidence in a grade if data has been collected at multiple times throughout the year compared to only once or twice throughout the year.

Locations of sampling sites that contribute data to water quality, seagrass and coral scores for the Mackay-Whitsunday 2016 report card.

Other factors that influence data confidence relate to the method used to collect the data. For example, there is higher confidence in remote sensed data that includes regular ground truthing compared to remote sensed data with no ground truthing.

The method for calculating confidence scores is based on the Great Barrier Reef Water Quality Protection Plan report card method for communicating confidence. For more information on what makes up the confidence score see our  technical reports.

How can we improve data confidence?

Confidence scores can be increased in a number of ways including:

  • Increasing the number of locations that are monitored within existing programs.
  • Increasing the number of times throughout the year that monitoring is undertaken within existing programs.
  • Better aligning methods for monitoring the same indicator between existing programs.
  • Creation of new monitoring programs to expand on the pool of data for any given indicator.
  • Increasing ground truthing for indicators that use remotely sensed data.