The Carb Education Gap
Part 3 of a series on the meta-NPS. Part 2 showed the experts can’t agree on how healthy a carb food is. Bring in the public and a second gap opens: people rarely see a carb food the way the experts do, and they almost always read it as worse.
Consumers rated 20 carbohydrate foods for healthfulness; we compared their beliefs to the meta-NPS. Only two foods line up with the experts. The rest miss, and the miss has a direction: the public reads most carb foods as less healthy than even the experts say. That gap, food by food, is a map of where nutrition education has the most to do.
The short version. Part 3 of the Meta-NPS series. Published as Erndt-Marino & Ghirardelli, JANA 2026.
Parts 1 and 2 showed the experts cannot agree how healthy a carbohydrate food is. Bring in the public and a second, different gap opens.
- Almost nothing lines up. Consumers rated 20 carbohydrate foods for healthfulness against a pooled expert consensus. Seventeen of the twenty land at Alignment Grade C or D. Only an apple and a soda actually match.
- The miss has a direction. The public reads most carb foods as less healthy than the experts do — every potato food, including a plain baked potato. Honey runs the other way, read as far healthier. The two juices invert: a sugary “vitamin C” drink earns a halo while genuine 100% juice gets marked down.
- And it is not simple ignorance. Seventy percent of the foods carry misperception in both directions at once. This is a structured set of beliefs, not a deficit to be filled by saying more.
- Where to start. Filter by misalignment, then by how much America eats it, then by whether experts are also unsure, and four foods survive: white rice, white bread, corn tortilla, and vitamin-C juice.
The argument across three parts. Whether we agree what healthy means (Part 2) and whether we communicate it clearly (this part) are both features of the informational environment. Most nutrition effort goes to changing behavior and access, on the assumption the information is settled. It is not.
Part 2 ended on a hard fact: run five expert systems across the carbohydrate supply and they disagree on most foods, grains worst of all. That is a problem inside the science. This part asks what happens outside it. If the experts are unsure, what does the public believe, and do the two ever meet?
To find out, NORC ran consumer focus groups on 20 carbohydrate-rich foods. People rated each one from less to more healthy, then said how certain they were. We lined those beliefs up against the meta-NPS — a single harmonized healthfulness score built from five expert rating systems, the same one from Parts 1 and 2. The comparison is preliminary and the limits from the earlier parts still apply: no method can prove which rating is correct, so this measures distance between beliefs, not who is right.
What a misperception is
Before the results, one simple idea. Put healthfulness on a line and mark where the experts (the meta-NPS) land. Someone who reads a food as less healthy than the experts sits to the left — that gap is negative perception (purple). Someone who reads it as more healthy sits to the right — positive perception (orange). Tally those gaps across everyone and you get the food’s total misperception, and a single food can pull both ways at once.
For whole-wheat bread most people land to the left of the experts, so the purple side outweighs the orange. Those two colours hold for the rest of the piece, and that two-part measure, sized into a single circle, is how every food gets summarised.
The gap has a direction
Run the comparison across all twenty foods and almost none line up with the experts. Each food earns an Alignment Grade — a school-style grade on the size of its consumer–expert gap, A (closely aligned) through D (far off), regardless of which way it leans. Seventeen of the twenty land at C or D; only three even reach an A or B, and just two — an apple and a soda — actually line up. (The third, a smoothie, earns a B but, as we will see, hides a real gap in both directions at once.) Put each food’s gap on one readout, then split it around the experts, and the shape of the misperception appears: which foods fall to the less healthy side, which to more healthy, and how many reach both ways at once.
The gaps lean one way. Bar after bar runs toward less healthy than the experts: the public is consistently harder on carbohydrate foods than the people who study them. Every potato food shows it — French fries, chips, and even a plain baked potato are all read as less healthy than the experts rate them, with air-popped popcorn close behind. The loudest gap of all runs the other way: honey, read as far healthier than the experts. And the two juices flip — a sugary “vitamin C” fruit drink earns a health halo while genuine 100% juice gets marked down. But the deeper pattern is that most foods are not one-directional at all: seventy percent carry negative and positive misperception at once — white rice splits almost evenly, and even the smoothie carries a real dose of each. Only an apple and a soda are read close to the experts. So the gap is not ignorance in one direction. It is a structured, food-specific set of beliefs — a map of exactly where the message has to work.
And the experts are split too
It would be easy to read the last chart as the public getting carbs wrong against a settled expert answer. There is no settled answer. These are the same 20 foods, now showing only the five expert systems and how far apart they land.
Fifteen of the twenty earn the meta-NPS verdict uncertain — the label the harmonized score gives a food when its five expert systems can’t agree, their ratings strung across the scale. Only a handful are settled: apple and popcorn high, soda low. So two gaps stack on top of each other. The public disagrees with the experts, and the experts disagree with one another. That doubled uncertainty, not a simple knowledge deficit, is the carb education gap.
Where to start
With twenty foods and two kinds of gap, the question becomes where to put limited effort. The lever, as in Part 1, is consumption: a contested, misperceived food barely matters if no one eats it. Filter the list down by the gaps and by how much America actually eats each food. Watch what survives — and what does not. A big public gap is not enough on its own: French fries are badly misperceived, but the experts mostly agree they are middling, so they drop out when the last filter keeps only the foods the experts are also unsure about.
Keep the misaligned foods, then the most-eaten, then the ones the experts are also unsure about, and four rise out of the twenty: white rice, white bread, corn tortilla, and vitamin-C juice. Each is misperceived by the public, eaten in quantity, and contested among experts. If messaging has to start somewhere, it starts there. French fries make the point in reverse: a loud public gap, but because the experts agree on them, the fix is plain communication, not harmonization. Different gaps call for different fixes.
Where this goes
Three parts, one argument. We do not have a single, agreed definition of a healthy carbohydrate food (Part 2), and the public’s beliefs diverge from the experts’ in a measurable, mostly pessimistic pattern (this part). Both of those are features of the informational environment: whether we agree on what healthy means, and whether we communicate it clearly. Most nutrition effort goes to the other two environments, changing behavior and changing food access, on the assumption that the information is settled. It is not. Closing the expert gap (harmonization) and closing the public gap (communication) is lower-hanging and high-leverage, and it feeds the rest: clearer definitions make better labels and better messages.
Two honest limits carry over from the earlier parts. Information enables better choices, it does not guarantee them; healthy foods are not the same as healthy diets. And every rating here, expert or public, scores one food the same way for everyone, when healthfulness also depends on the eater. Harmonizing the universal disagreement is the floor; personalizing beyond it is the ceiling. This series has been about measuring the floor honestly, so the gap is visible before anyone tries to close it.
The 20 foods, the alignment grades, the misperception measures, and the four priorities are from the public IAFNS Alignment Between Experts and Consumers webinar and the companion analysis (Erndt-Marino & Ghirardelli, on educational gaps in carbohydrate healthfulness). Consumer ratings came from NORC focus-group screeners; expert ratings are the five-system meta-NPS. The numbers are a conservative floor: only five of hundreds of published rating systems were sampled, and a single set of cutoffs drew the lines. The webinars and papers behind the series are on the resources page.
Comments