Nutrition data is a complex beast and we often find people misunderstand some important limitations of our nutrient database that can lead to misinterpreting the results we show. To help make better sense of things, we've introduced a new statistic called Data Confidence that you can see for any reported nutrient target value, or collectively for all of your targets.
The key issue is that while we allow you to track over 70 different nutrition data points, not every food in our database has a data point for all 70 nutrients that we display. When you look at your daily total potassium intake, for example, this is the sum of all the potassium from all of the foods you've logged. However, if some of the foods you've logged contain potassium, but the database doesn't have a value for potassium for that food item, then your total for the day will be undercounting that nutrient.
This happens, because many items in our database, in particular branded, packaged, food products and chain restaurant menu items do not typically have data available for all 70 nutrients that we track. On American food labels, only about 15 or so nutrients are usually reported (the minimum required by law), and often, that is all that is available.
Data Confidence looks at each nutrient value in each food and based on the type of food and data source, applies a score to help you determine how 'complete' your data for that nutrient is.
For example, here are two different Quaker Chewy Granola Bar products in our database. On the left, we have "Quaker, Chewy Granola Bar, Peanut Butter Chocolate Chip" which is an entry from our highest quality data source (NCCDB). It's been fully analyzed and has 76 listed nutrients. To help you spot these high-quality data items, we use the red laboratory beaker icon. On the right is a similar granola bar "Quaker, Chewy Granola Bars, Raspberry Fruit Crumble" but we only have 16 listed nutrients for this product because it hasn't been fully analyzed by the NCCDB. Instead, all we have is the 16 nutrients listed on the product's packaging. Because this is from a product label, we use the barcode icon to make it easy to spot.
If I log both of these foods to my diary and then examine some of my nutrient targets, I can now see how this impacts the data confidence scores. If we look at Potassium, I can see the data confidence score is 100% because both foods contain data for Potassium.
But if I look at Magnesium, the confidence for our magnesium total is only 50% because we are missing data for Magnesium in one of the granola bars.
In the nutrition scores popups you can also see the data confidence for the set of nutrients examined by that score. For example here, at the bottom of the display, we show a score of 84.4% which is the average confidence of Folate, Vitamins A, B12, C, & D, Copper, and Zinc.
We've also added a 'Data Quality' score that lets you show the overall data confidence for your complete set of tracked nutrient targets.
You can even chart it on the Trends -> Charts tab and get a picture of how comprehensive your data is overtime!