How Cooking Method Changes Nutrient Values (and Why Raw ≠ Cooked)

Cooking doesn't destroy nutrition evenly — it reshapes it. In FoodData Explorer's own measured values, boiling cuts kale's vitamin C by 81% and its potassium by nearly 59%, while protein barely moves. USDA tracks this with a 270-entry retention-factor table precisely because nutrient loss depends on the food, the nutrient, and the cooking method.

---

What actually happens to a nutrient when you cook it?

Three things can happen to a nutrient during cooking, and they're easy to conflate: it can leach into cooking water or drippings and get discarded, it can degrade chemically from heat or oxygen exposure, or it can simply get concentrated as water evaporates and the food shrinks. Only the first two are "losses" in any meaningful sense — the third can make a cooked food look more nutrient-dense per 100 grams than its raw counterpart, even though the total amount of the nutrient hasn't changed. This is why a blanket claim like "cooking destroys nutrients" doesn't hold up against real measured data: the direction and size of the change depends on which of these three mechanisms dominates for a given nutrient in a given food.

How much does cooking actually change a real vegetable's nutrient profile?

FoodData Explorer's database has a genuine same-food pair that shows this directly: kale, raw against kale, frozen, cooked, boiled, drained, without salt, both measured per 100 grams. The differences are large and uneven across nutrients:

• Vitamin C: 93.4 mg raw → 17.8 mg cooked — an 81% drop
• Potassium: 348 mg → 144 mg — a 59% drop
• Sodium: 53 mg → 16 mg — a 70% drop
• Thiamin: 0.113 mg → 0.063 mg — a 44% drop
• Calcium: 254 mg → 150 mg — a 41% drop
• Iron: 1.6 mg → 0.84 mg — a 48% drop
• Vitamin A, RAE: 241 mcg → 146 mcg — a 39% drop
• Magnesium: 32.7 mg → 25 mg — a 24% drop
• Zinc: 0.39 mg → 0.27 mg — a 31% drop
• Protein: 2.92 g → 2.94 g — effectively unchanged

One honest caveat on this pair: the cooked sample is frozen-then-boiled, not fresh-boiled, so blanching before freezing contributes to some of the loss alongside the boiling step itself — it isn't a perfectly isolated single-variable comparison, but it's the real measured pair this database has, and the pattern (water-soluble vitamins and minerals drop hardest, protein barely moves) matches what food science expects of a boiling-type method. MedlinePlus confirms the general mechanism directly: cooking vitamin-C-rich foods, or storing them too long, reduces their vitamin C content, and the best food sources of vitamin C are "uncooked or raw fruits and vegetables" specifically because the vitamin is both heat-sensitive and water-soluble.

Why did potassium and sodium drop by more than half if they aren't "heat-sensitive" vitamins?

This is the leaching mechanism at work, and it's a different story from vitamin degradation. Potassium and sodium are minerals — heat doesn't destroy them the way it can break down an ascorbic acid molecule — but both are highly water-soluble, so when kale sits in boiling water, a meaningful share of its mineral content dissolves out into that water and leaves with it when the food is drained. The kale data shows this clearly: sodium fell by 70% and potassium by 59%, both larger percentage drops than iron (48%) or calcium (41%), even though none of these four minerals is "destroyed" by heat in the way a vitamin can be. The size gap between the mineral losses is itself informative — sodium and potassium are freely dissolved ions in the plant's cell fluid, so they wash out fastest, while calcium and iron are more often bound into cell-wall and protein structures that resist dissolving into the cooking water quite as readily, which is consistent with why their percentage losses, while still substantial, came in lower. The practical takeaway is method-specific: a cooking method that keeps the food out of standing water (steaming, roasting, stir-frying) will leach far less mineral content than one that submerges the food and discards the liquid, because the loss pathway here is physical dissolution, not a chemical reaction with heat itself.

Why does USDA track 270 different "retention factor" categories instead of one rule?

Because the answer genuinely depends on the food and the method, not just "cooked vs. raw." FoodData Explorer's own retention_factor table — part of the same USDA data system this site is built on — has 270 rows, each describing one specific food-group-and-preparation combination: "CHICKEN, BROILED," "VEG, GREENS, BOILED, WATER COVER DRAIN," "POTATOES, BAKED IN SKIN," "FINFISH, <5%FAT, FRIED, W/COATING," "LEGUMES, CKD 45/75MIN, BOILED, DRND," and so on across 13 distinct food-group categories (cheese and eggs, poultry, grains/cereal, fruit, pork/bacon, vegetables and potatoes, nuts, beef, alcoholic beverages, finfish, legumes, lamb, and flour/meal). The sheer granularity is the point: USDA's own nutrient scientists concluded that "boiled" doesn't predict nutrient retention consistently enough across a chicken breast, a potato, and a leafy green to be treated as one factor — each food-group-and-method pairing gets its own entry. This table is the USDA Table of Nutrient Retention Factors methodology distilled into this database's own structure — a system built by USDA's Nutrient Data Laboratory specifically to model how cooking changes the nutrient content of roughly 290 foods across 16 vitamins, 8 minerals, and alcohol, method by method, rather than assuming one universal retention percentage.

Does that mean every cooked nutrient value in this database is a calculated estimate, not a direct measurement?

No — and this is a distinction worth understanding if you're reading values closely. FoodData Explorer's food_nutrient_derivation table lists 64 separate codes describing how each published nutrient value was actually derived, and only a handful of those codes explicitly reference retention factors at all: 5 codes are documented as using retention factors in a concentration adjustment (when a value is calculated from a related food form) and 12 are documented as explicitly not using them. The rest of the 64 codes describe entirely different derivation paths — direct analytical measurement, label-claim back-calculation, recipe-based calculation, summed components, and more. In plain terms: some published figures are direct lab measurements of the actual cooked food, and some are calculated from a related food's measured value using a retention-factor adjustment — the derivation code attached to each figure tells you which. The kale pair cited above carries no derivation code at all in this database, meaning both the raw and cooked values are standalone measured entries, not one calculated from the other.

Does cooking ever increase a nutrient's measured value instead of lowering it?

Yes, and the kale data demonstrates the mechanism even without an increase in this specific pair: protein held essentially flat (2.92 g raw → 2.94 g cooked) despite every water-soluble vitamin and mineral measured falling, because protein isn't lost into the cooking water or broken down by heat the way vitamin C is — so as water content changes during cooking, protein's share of the remaining mass stays comparatively stable while water-soluble nutrients around it decline. This same water-loss concentration effect is why heat itself isn't automatically the enemy: a method that drives off water without leaching the food in standing liquid (roasting, grilling, dry sautéing) can leave non-water-soluble components looking unchanged or even higher per 100 grams, while a boiling-and-draining method on the same food tends to show the sharpest mineral and vitamin-C losses, as it did here. This is also why a packaged food's nutrition label can sometimes show a higher per-serving mineral or protein figure after cooking than a shopper expects from the raw product alone — the food has simply lost water weight, concentrating whatever didn't leach out or degrade into a smaller remaining mass, not gained any new nutritional content from the cooking process itself.

So is raw always the better choice?

Not based on what this data shows. The evidence here points to a method-and-nutrient-specific answer, not a universal one: if the priority is preserving water-soluble vitamin C and leachable minerals like potassium, a method that avoids submerging food in water that gets discarded afterward will measurably outperform boiling-and-draining, regardless of whether the food is "cooked" in some other sense. If the nutrient in question is a protein or a non-water-soluble component, cooking method has far less practical effect on the measured value. The honest, data-grounded answer is to treat "raw vs. cooked" as a question about the specific nutrient and the specific method — which is exactly why USDA built a 270-category retention-factor system instead of a single raw-to-cooked conversion rule.

---