In this LAB note
Question · Hypothesis · Method · Results · What Failed · What We Learned · What Changed at MOMS
Question
When a MOMS shake pours 137 ml more liquid, what changes besides its weight: the protein, the calories, or both?
Hypothesis
We expected protein to stay put and weight to rise. Protein in our recipes is set by the protein slot (we hold a 30 g floor, see LAB #028), and liquid is poured after that decision, so a shake with more liquid should be heavier and less energy-dense but carry the same protein. We expected calories to fall only slightly, if at all, because the extra volume is mostly plant milk or coconut water, which add few calories per ml. The published energy-density literature predicts the same direction: lowering a food's energy density by adding water lowers the energy eaten later, without changing what the food supplies.
Method
Published research: Bell et al. (1998, Am J Clin Nutr) fed normal-weight women meals of low, medium, and high energy density; they ate a similar weight of food in each condition, and their energy intake rose with density (about 1376, 1519, and 1800 kcal). Rolls, Bell and Thorwart (1999) found that water incorporated into a food lowered later energy intake, while the same water drunk alongside the food did not. Internal analysis: we took all 201 recipes in the MOMS database and summed each recipe's listed liquid in ml. We compared the 93 recipes with 100 to 200 ml of liquid (mean 164 ml) against the 78 with 280 to 350 ml (mean 301 ml); the 30 in between, and one recipe with no liquid, are left out. Shake weight is the sum of every listed ingredient in grams or ml, counting 1 ml of liquid as 1 g; it is not a measured volume of the blended drink. We then checked the gap inside one liquid type (coconut milk only), inside each Cal Label, and with a simple regression of calories on liquid ml and grams of legume in the protein slot. Two recipes are shown by ID: EY-W1-ENERGY-009-V1 (Dawn Rush) and EY-W2-DETOX-002-V1 (Soft Reset).
Results
Protein did not move: 33.1 g in both groups. Everything else did. Average shake weight rose from 527 g to 667 g (+140 g), carbohydrate fell from 43.2 g to 34.4 g (-8.8 g), and calories fell from 386.1 to 362.1 (-24 kcal, -6%). Fiber slipped from 7.9 g to 7.4 g. Because weight rose while calories fell, energy density dropped from about 0.73 kcal per gram to about 0.54, roughly 26% lower, with protein unchanged. Across all 171 recipes the correlation between liquid ml and protein is -0.03; between liquid ml and calories it is -0.38. The gap holds up under two checks. Inside each Cal Label the high-liquid recipes are lighter by 17 to 20 kcal (Filling 388 vs 369, Standard 383 vs 366, Light 374 vs 354). Inside coconut milk alone (20 low-liquid and 22 high-liquid recipes) calories are 391.7 vs 366.3, a 25 kcal gap, with protein at 34.2 vs 33.6 g. In the regression, each additional 100 ml of liquid goes with about 14 fewer kcal after accounting for legume grams, and each additional 50 g of legume in the protein slot goes with about 6 more kcal.
What Failed
The mechanism we expected did not hold up, and one pair of recipes shows it. Dawn Rush (EY-W1-ENERGY-009-V1) pours 150 ml of coconut milk and weighs 504 g; it reads 386.4 kcal, 30.9 g protein, 33.0 g carbohydrate. Soft Reset (EY-W2-DETOX-002-V1) pours 300 ml of the same coconut milk and weighs 696 g (+192 g); it reads 389.0 kcal, 32.8 g protein, 32.3 g carbohydrate. Doubling the liquid here added 2.6 kcal, not removed any. The two recipes also differ elsewhere: the egg white rises from 150 to 220 g and the chickpeas fall from 60 to 30 g, so the liquid was never changed in isolation. The group-level gap is therefore not a per-recipe dilution effect. Low-liquid shakes lean on oat and rice milk (46 of 93 recipes) while high-liquid shakes lean on coconut, almond and soy; 15 of 20 low-liquid coconut-milk shakes put a legume in the protein slot against 8 of 22 high-liquid ones. We tried to explain the gap by liquid type and by legume grams, and neither accounts for it: the coconut-only gap stays at about 25 kcal. Light recipes are also over-represented in the high-liquid group (28 of 78 against 6 of 93), so the Cal Label is a competing explanation even though the gap persists within each label. The liquid volumes are formulation choices, not an experiment, so this is an association in 171 recipes and not a measured effect of liquid.
What We Learned
In our recipes, liquid volume is the loosest variable in the formulation: it moves the weight of the shake by roughly 140 g and the protein by nothing. Protein is decided in the protein slot, and a shake that reads as larger is not a shake that carries more protein or more food. The calorie difference that travels with more liquid is real in the data but we cannot assign it to the liquid itself; it travels with a different mix of liquid types, legume doses and Cal Labels. What the literature adds is the reason the weight matters at all: when the amount people eat stays roughly constant, a lower energy density means fewer calories eaten.
What Changed at MOMS
We now set the protein slot first and treat liquid volume as the last variable we adjust, not as a lever for protein or for the Light label. When a recipe moves between about 150 ml and about 300 ml of liquid, we re-check its protein and its calories against the label instead of assuming they hold. And we no longer describe a higher-liquid shake as lighter on the strength of its volume alone; if a recipe needs to read lighter, we will look first at the carbohydrate sources in the body and protein slots.
References
Bell EA, Castellanos VH, Pelkman CL, Thorwart ML, Rolls BJ. Energy density of foods affects energy intake in normal-weight women. Am J Clin Nutr. 1998;67(3):412-420. | Rolls BJ, Bell EA, Thorwart ML. Water incorporated into a food but not served with a food decreases energy intake in lean women. Am J Clin Nutr. 1999;70(4):448-455.
Keep reading
Previous experiment
Not sure which shake fits you? We match the formulation to your body type.
Find your body type