Among the many problems in the extended “dumbing down” and/or “simplistic thinking” families is the unwarranted reduction of complicated and multidimensional matters to a single dimension, often in form of some scale—as with the Nutri-Score baloney discussed below. Longer discussions of political scales are present in several texts, most notably in [1], while at least one text deals with partially overlapping ideas in forms of spectra and a spectrum fallacy ([2]).
Whether “dumbing down”, “simplistic thinking” (an ad-hoc term), or something else yet applies depends on the details. Nutri-Score seems to be dumbing down, as it is something told by self-appointed experts to expected-to-be-dumb-as-doornails consumers. However, the deeply flawed Left–Right scale is often a case of simplistic thinking by those who, themselves, do not understand better, and the same likely applies to many cases discussed in [2].
To make matters worse, issues like propaganda angles cannot be ruled out, e.g. in that the German Left uses the Left–Right scale to (with great and lamentable success) mis-identify anything “Rightwing” with Nazism, racism, and similar, making the label unusable for Libertarians and Conservatives, creating great guilt-by-association issues, and putting anyone not explicitly Leftwing at risk of being condemned as “Nazi”, “racist”, whatnot. The more the shame, as the Nazis were Leftwing (cf. [1]), as racism has nothing to do with the matter and is currently far more common on the political Left, etc.
Even with Nutri-Score (cf. below) there is a possibility that there is an angle of deliberate manipulation involved—not just plain dumbing down.
A recent/2025 annoyance in this direction is the currently ever present Nutri-Score labels displayed prominently on groceries in Germany, which presume to give an A–E, single scale, verdict as to whether a particular food is healthy or unhealthy and (by implication, after factoring in price) a good or poor purchase. However, chances are that these labels do more harm than good or, on the outside, that what good they do could be achieved better and with a greater net benefit by some other means. Consider (in a likely very incomplete analysis):
There are too many relevant dimensions for this to be a good idea. I might suggest an alternative of four independent scales that indicate respectively how energy dense a food is, how highly processed it is, how “nutritionally positive” it is in terms of e.g. micro-nutrients, and how much junk it might contain (e.g. corn syrup, refined sugars, excess salt, ...)—and even this would be a very great over-simplification.
But an over-simplification that could give a reasonable trade-off in terms of understandability vs. informed decision—the more so as further information is still usually available in a tabular “table of contents” on the back of food packages. (And making these mandatory and more prominent would bring the thinking customers more value than the Nutri-Score labels.)
The actual Nutri-Score calculations seem to go by a mixture of what is “nutritionally positive” and what is junk, while ignoring both energy density and level of processing—and to contain a number of exceptions and other complications to boot. (Also see the next item for some oddities.)
Looking at actual scores, there are plenty of oddities.
For instance, I recently bought some very white, very processed, toast bread—something usually viewed as broadly negative. (And at least partially for reasons that Nutri-Score does not seem to consider, notably, the apparent risk of blood-sugar spikes. Also see a below side-note.) However, this bread carried a proud “A”—the highest grade. But assume, for the sake of argument, that this “A” was fair: Where does that leave bread normally considered healthy? Not one iota better, because there is no further differentiation into, say, “A+”, “A”, and “A-”.
For instance, I recently bought two pizzas from the same brand that appeared identical except for one thing: the one had a mushroom topping, the other a salami topping. The mushroom one scored a “B”; the salami one, a “D”. A switch merely from mushroom to salami, making up only a small proportion of the overall of the pizza, would then justify a drop of two “grades”. Something is not right here. (With reservations, of course, for any hidden differences beyond the topping, but I noticed none during the respective eating.)
For instance, I had a look at (but did not buy) a jar of honey—something normally considered healthy. Score? Either a “D” or an “E”. (My memory fails me.) Now, not all honey is created equal, up to the point that there seems to be occasional cases of fake honey, but here we might have a failure to consider typical use. I could imagine that honey would be unhealthy if someone takes a Winnie-the-Pooh approach and empties a big jar in a single sitting—but that is hardly how honey is normally used. More typical uses include as a better-than-white-sugar sweetener applied in small doses and as a home remedy or ingredient in a home remedy. (I leave unstated whether such home remedies work—the point is that different products cannot be compared fairly without considering actual intended use.)
I skimmed through a Swedish text on such fake honey earlier in 2025, but I am not aware of how common the problem is and whether the problem exists in Germany too. The idea would be to take sugar syrups from other sources, add some extras to make the result “honey-er”, and then to apply a false label. (I do know that similar problems can occur with other products in Germany, e.g. in that the label “wasabi” usually implies “horse radish with some green coloring”.)
As can be seen from the previous item, there is a problem with granularity. If e.g. most breads are awarded an “A”, how do we differ between “good” and “bad bread”? In reverse, why not pick a scale of 1–100? (Or 0–100. The best choice for lowest score is debatable. Note that sufficient information is available before the Nutri-Score calculations are normalized to those A–E values and/or can be made available by adapting the calculations slightly.) This the more so as the true difference between an “A” and a “B” can be small and the difference between two “Bs” can be large. Say, for the sake of argument, that values of 85, 84, and 71 on a 1–100 scale are mapped to “A”, “B”, and “B” on the “A–E” scale. Certainly, the first “B” has more in common with the “A” than with the second “B”.
Such lack of granularity could partially explain the above pizza issue, if the mushroom pizza had a very weak “B” and the salami one a very strong “D”; however, this still leaves the entire “C” range to be explained by the change in topping. In terms of distortions of customer perception, it might even be worse than if the true distance was closer to the nominal two steps.
Any such simplistic scale invites manipulation and attempts to “game the system”. (This even outright cheating aside, as with the emission-measurement scandal around Volkswagen in 2015.) Looking back at the previous item, the maker of the first “B” would be well advised to improve his score from 84 to 85 (a trivial step) to earn an “A”—but that improvement in letter scores does not in the slightest reflect the underlying improvement. Going with several scales or e.g. a 1–100 scale would reduce the scope for such gaming considerably.
Such scales fail to consider differences in preferences, use (cf. honey above), takes on what is or is not “good” and “bad” in terms of food (where even official recommendations are known to change considerably over time; cf. parts of a text on bad advice), how a food fits within an overall diet, and similar. Consider e.g. how perceived “goodness” and “badness” is more a matter of having too much or too little of something than a matter of some absolute characteristic, with consequences like a salt-rich food being a negative in an already overly salt-rich diet but potentially a positive in a salt-poor one; or how level of physical activity and training goals can affect what is or is not a sufficient quantity of this-and-that.
Such complications have the interesting side-effect that techniques like principal component analysis (otherwise seemingly good choices to address such mappings of multiple dimensions onto one) are less likely to be helpful than in more clear-cut settings. (And even within principal component analysis it is far from certain that any given setting allows a reasonable reduction to a single component. More often, the realistic aim is merely to have significantly fewer components than originally.)
Great unclarity exists as to what comparisons Nutri-Score allows. The obvious and intuitive case is to compare different products and/or product categories, with an eye at an overall dietary improvement, but, allegedly, the opposite of comparisons within a product category is intended. This, however, is not in the slightest clear from the labels and great problems with e.g. granularity are likely to arise, as with the bread above. Contrast this, again, with multiple scales and/or a higher granularity, which would allow both comparisons between and within product categories with far less ambiguity.
An additional ambiguity arises through what is or is not meant by a phrasing like “product category”. (And I admit to not having done the legwork with regard to the mentioned Nutri-Score intents.) Do we e.g. have apples vs. apples but not apples vs. oranges or do we have apples vs. oranges but not apples vs. bread?
Either way, the situation is unfortunate.
(And if a scale does not allow reasonable comparisons between, say, bread and soft drinks, well, that is hardly worth mentioning, because they are so different in character, typical way of consumption, typical quantities consumed, etc.)
Manipulation by the Nutri-Score creators cannot be ruled out, e.g. in that criteria or weights of criteria are chosen to favor or disfavor various types of foods for illegitimate reasons. For instance, if red meat scores poorly this could be because red meat is often considered a less healthy type of food (but note an earlier item), but it might also go back to recent pushing of “Red meat causes global warming!!! We must never, ever eat read meat again!!!”. Likewise, that surprising “A” for white toast bread might go back to an objective assessment, but it also fits a prior pattern of official recommendations (food pyramids and whatnots) pushing for more bread than many consider healthy. (In both cases, note that I point to the possibility without making a definite claim. If such manipulation takes place, it is worse, but even the potential of manipulation is bad in its own right, because it reduces the credibility of such scales.)
Even absent deliberate manipulation, criteria and weights of criteria (more generally, combination of criteria, which can be more sophisticated than in this example) can have a very great effect on relative rankings. For an easily understandable example, assume that someone creates “size scales” for humans that use height but not mass, mass but not height, or both in varying proportions. Indeed, even something as simple as just adding the numerical value of height and mass (for a trivial calculation) can lead to very different results if units of e.g. centimeters and kilograms, inches and stones, feet and pound are used. It might seem that height and mass are weighted equally by the straight addition, but there is actually an implicit weighting through the choice of units. Likewise, assume that someone attempts to reduce the (off topic) distortion caused by the correlation of height and mass by replacing one of the two with some other measure (e.g. BMI): Firstly, the choice of such measure is somewhat arbitrary and the choice can affect outcomes strongly. Secondly, even if a satisfactory measure is found, the problem of weighting remains.
Above, I use “mass” in part to avoid confusion with “weight” in the sense of relative importance when combining criteria. (In part, because it is more accurate through being less dependent on local gravity and whatnot.)
Issues around units can give further hints about what types of combinations are unsuitable, e.g. in that addition of quantities that are too different in nature should raise warning bells. Contrast e.g. adding two distances in the same unit with adding distances in centimeters and distances in inches, distances in centimeters and masses in kilograms, and (for many purposes) grams of fat and grams of protein.
On a more meta-level, Nutri-Score could in part be considered redundant in that it very often matches a heuristic of “kids like to eat it; ergo, low score” and “kids do not like to eat it; ergo, high score”, maybe, combined with some few other, easily internalized, heuristics, e.g. that red meat tends to lead to a low score.
Some surprises do exist, as with the white toast bread above—but to what degree can such surprise classifications actually be trusted? (Consider e.g. that blood-sugar spike: Is Nutri-Score naive or are the often voiced concerns in that area overblown?)