I am a “weight rather than volume” person for flour and other “pouring” powders. This surely is, because this is normal and everyone does it this way and always had. My grandma’s cookbooks already are full of weight measures for things such as flour in baking. Everyone has done it forever – everyone, always, forever? Nope. One of the perks of moving and travelling internationally is, you learn EVERYTHING can be done differently and thought to be the perfect standard usual norm, which applies everwhere. My parents discovered this in Scandinavia in an amusing way, where they bought cherries by the liter. Initially thinking they misunderstood, then showing rock-solid old-cartoon-style question marks right in front of their foreheads, they measured them with a litre-measure. You find it amusing that someone things it’s weird? Welcome to the planet, things are done differently in different places and your expectations can be highly amusing – and puzzling. But back to the subject!
Why even bother?
The reason in cooking and baking to have a recipe and measure ingredients is to get consistent results that are reproducible at different times and in different places.
One of my most related baking adventures for this topic has been making my very own Luxemburgerli. I missed them here in Australia and took up the challenge to make something similar. They are like macarons, with a butter-cream filling and there is no way you get them right, if you don’t start with the right consistency of the dough (I will post the recipe at some stage).
Liquid(ish) powders – e.g. “flours”
Those are the most problematic in my view. Weighing and also volume measurement of powders are used widely, to get the “right” amount of the ingredient. I will only reason about (wheat) flour here, because in cooking and baking it is one of the most common ones and serves as a good example (plus gathering details about all sorts of common edible powders really is a PhD in itself).
In your recipes, a reproducible amount of “pure absolute dry flour” (without any absorbed moisture, I will call it “dry flour” in this post) is the quantity you are interested in to get your recipes to work repeatedly. This aim needs to be achieved with household tools and without major hassle. However, flour(s) and other powders contain a certain percentage of bound water, depending on environment temperature and humidity and the specific water-holding properties of the powder. Thus you cannot measure this dry flour weight directly. Also, the changing moisture in the flour prevents you from assuming a simple conversion formula. On the other hand, the volume in a certain “fluffyness” does not change with the (small amounts of) moisture much, but reproducing a certain degree of fluffyness is nigh impossible, since every handling changes the amount of “air volume” in the flour. So both have flaws, but which one is more reliable?
Weight Error due to Moisture
Standardizing the moisture content of your flour requires highly conditioning your kitchen and storage area in terms of air moisture and temperature. You would also have to make sure you settle your ingredients for a certain amount of time to equilibrate in the new environment. Not many would be prepared to go that far, I certainly am not and thus tried to figure whether the error introduced by moisture differences is big(ger than in volume measurement).
I found the publicly accessible book of Harry Snyder (1922), which addresses the moisture content problem in terms of trade regulations. It basically states that it may dry to 9% of moisture (or less) and up to 15 % (or more). Wheatflourbook states that above 14 % of moisture in flour, it starts molding easily, which I assume most people would notice and buy fresh flour from the store, stored at conditions that keep the flour at sufficiently less than that. Californiawheat state for semolina (coarse wheat milling product), that it usually ranges between 12 and 14 % of moisture and I’ll take that as a proxy for flour in a cupboard. There is a sampling error as well mentioned, which tells you that the moisture is not evenly distributed in even one pack of flour.
However, let’s assume moisture ranges around 9-15 % in general wheat flour, if stored in a household, where room temperatures and moisture are not extreme. This gives me roughly a 15%-9%=4% weight fluctuation range. Four grams per 100 grams, my kitchen scale can do that – roughly, just.
Volume Error due to Fluffyness
Now this was tricky. Fluffyness is not exactly the right scientific term to search for, despite its obvious advantages to understand the “state” in which the flour is in (*hehe). Given that the moisture content does not influence the volume much, you may check on the following yourself. King Arthur Flour from the US gives an example of their own struggle to reproduce volumetric amounts, stating originally a cup for them was 113 g (4 ounces). You can, according to them, in fact make a (I assume U.S.) cup of flour 113 g, by sifting the flour first and thus incorporating air-volume into the flour (this is somewhat like the moisture-weigh you measure in the weighing approach). If you simply scoop it, you can get 156 g and more (5.5 ounces). So unless you absolutely clearly state how the flour is treated (reproducibly of course) and you are extremely careful to not change that volume by much handling prior to measuring, you may end up with anything between 113 g and 156 g per cup. This is 38 % or more differenc in flour amount. I guess this is obviously way more than the estimated 4 % difference when weighing.
Liquids and solids
In general, liquids and solids such as milk and water, and butter and nuts undergo little variation in volume and weight due to the environment in a normal household. While milk and water fill out the measuring device to the edges, solids don’t. Milk, water and other liquids are thus easily measured by volume, even though for water weight is just as appropriate (1 kg per 1 L) and a lot of water based liquids such as milk are so close to water in density that you can generally safely weigh them (1.03 kg per 1 L for milk). Oil based liquids often are more different in density (e.g. sunflower oil – 0.920 kg per 1 L). Make your call depending on how accurate you need the measurement to be and also how accurate your measuring device is. You may not get a good accuracy out of a two litre measuring jug for measuring 50 mL. Be watchful of the parallax error as well!
Accurately measuring the volume of solid butter, chocolate and nuts is difficult, because they don’t fill out the measuring device entirely, leaving an unknown amount of air-space-volume free. You could fill the measuring device up with water and then measure the volume of water and subtract, or liquefy butter and chocolate (and toss out what you don’t need). This is, however, laborious and changes your ingredient state or makes them wet. Consequently, weight is the most accurate and convenient way of measuring these in most cases. That said, if the accuracy in amount (e.g. weight) doesn’t matter much and the recipe is sufficiently accurate (finely chopped, coarsely chopped walnut), there is no reason to be overly precise and complicate things.
So let’s look into what consistency in results you get from weight and volume measurements for “fluid” powders, where it may be important and which one you may want to use in which context. Sure enough, the conclusion will be my personal one, but giving you the information allows you to choose for yourself.
Other errors and problems
Errors in measuring due to “wrong execution” in volume are much more frequent and easier to make than with scales. Think of the parallax error, think of the exactness with which you have to fill your measuring cups to actually get an exact cup (and which one for this matter – U.S./metric), half-cup etc. How often you may spill said cup, because it is literally filled over the brim. Think of overfilling and making things messy around the container to measure your exact 1/3 cup. Think of the room for interpretation when using “heaped”, “level”, “packed”. Thus, using a large container to measure weight is much easier and less messy in a lot of cases. Using scales, especially for anything above 5 g (a little depending on the accuracy of your scales), appears much more appealing and accurate to me in many cases when cooking and baking.
All that said, in cooking it usually does matter a lot less whether you have a teaspoon or two, 3/4 of a cup or a whole cup, while baking is much more texture-sensitive with respect to ratios. Furthermore, there are few recipes in home-amounts baking that only allow a very small margin of error. Most are a bit more flexible. However, I find that reproducible results especially in baking rely on the most reliable method to get your ratios right and that remains in most cases weighing (with something like 4 % of error range in flour) with a (good) kitchen scale compared to volume measurement (with something like 38 % of error range in flour).