Weighing a portion answers a quantity question. It does not, by itself, identify the food or analyze its nutrients. A useful record keeps the quantity and its nutrition reference visible as two separate inputs.
USDA FoodData Central contains different kinds of evidence: Foundation Foods uses analytically derived data for foods and samples, while Branded Foods uses manufacturer label information. The database name alone does not tell you how a particular value was obtained. [1]
Ask two questions before trusting the total
What did I measure or estimate? Identify the amount and unit: for example, “160 g weighed” or “one bowl estimated.”
What reference did I apply? Identify the actual food entry, package, or recipe and whether its description matches what you recorded.
Keeping those two answers together makes it easier to correct a record later, without reconstructing every decision from memory.
The same total can come from different inputs
This example uses an imaginary food reference. Its nutrient values illustrate the calculation and do not describe a real food.
Suppose your chosen reference says 8 g of protein per 100 g. A portion weighing 160 g gives a calculated 12.8 g of protein: 8 × 160 ÷ 100. Now imagine three different ways that entry could have been made.
Three records with different unresolved questions
Record | Quantity evidence | Reference evidence | What remains to check |
|---|---|---|---|
A | 160 g weighed | The intended food reference was selected | The reference describes the food; it does not analyze this particular portion |
B | 160 g estimated | The intended food reference was selected | The amount remains estimated |
C | 160 g weighed | A similarly named entry was selected without checking details | The food match remains unresolved |
All three might display 12.8 g after the same multiplication. The identical output does not erase their different inputs. Adding more decimal places would not resolve either the estimated portion in B or the unchecked food match in C.
Use the description before the number
In USDA Foundation Foods, descriptions include characteristics such as raw or cooked state, and nutrient values refer to edible material. Samples can also vary. Check those details before treating two similarly named entries as interchangeable. [2]
For your record, compare the food name, preparation description, included components, and quantity basis. If you weighed a prepared food but selected an entry described in a different state, pause and find a matching reference or explicitly record that the match is uncertain. Do not solve the mismatch by silently relabeling the weight.
The same principle applies to a saved recipe. A clear name such as “batch with added sauce” is more useful than silently reusing an older entry when the ingredients differ. You can choose how much detail to keep, but the description should not imply a distinction you did not verify.
Make a correction that explains itself
Return to record B. If you later establish that its portion was 120 g, the same fictional reference gives 9.6 g of protein. Keep the reason beside the change: quantity corrected from an estimate to a measured amount. The difference here is 3.2 g; that does not predict the error in another meal.
For record C, first resolve the food match. Changing the weight without resolving the reference would fix the wrong input. If you cannot establish a closer match, keep the uncertainty explicit rather than inventing precision.
Aim for a record you can understand later
A short note can be enough: which food reference you used, how you got the portion amount, and what still needs checking. If better information arrives, correct that part and keep the reason. Aim for a record you can understand when you return to it.
Sources & further reading
USDA FoodData Central: Data Type Documentation. Accessed 2026-09-28.
USDA FoodData Central: Foundation Foods Documentation. Accessed 2026-09-28.
Drafted with AI assistance and checked in a separate automated editorial review. This article has not received human or medical review. Our editorial standards.