Does it matter when I eat, or only how much?

What the experts say

Satchidananda PandaPhD · circadian biologist, Salk Institute (affiliation as of 27 Jul 2026; resignation effective 13 Aug 2026)

When is a separate lever

What it rests onMouse work from 2012 showing time-restricted feeding prevented diet-induced metabolic disease without reducing calories; a 2020 single-arm human pilot (n=19); a 2024 randomised trial in metabolic syndrome (n=108 completers) showing a 0.10 percentage point HbA1c advantage; and a 2024 pilot associating late eating with worse glucose tolerance after adjustment for body weight, fat mass, energy intake and diet composition.

Source: pmc.ncbi.nlm.nih.gov ↗

Layne NortonPhD · nutrition scientist, coach, and competitive powerlifter

How much, and timing is adherence

What it rests onControlled feeding trials in which total energy intake dominates timing effects.

Overview

They are measuring different outcomes in different people. Norton's endpoint is body composition in people trying to change it, and on that endpoint the calorie account dominates and he is on solid ground.

Panda's endpoints are glucose, blood pressure and lipids, often in people with metabolic syndrome who are already on statins and antihypertensives, and his strongest signals come from the subgroups whose baseline numbers were worst. Neither has shown the other wrong on the other's endpoint.

Where they agree

Both accept that a shorter window usually reduces intake without effort, and both treat that as a real benefit rather than a confound to be embarrassed about. Panda's own group concedes in print that most trials did not measure when people actually ate, which leaves Norton's incidental-reduction reading open on much of the literature. Neither recommends grazing.

What would settle it

A trial with calories genuinely equated, eating times actually measured rather than assumed, and metabolic endpoints alongside body composition. Panda's 2022 review names the missing measurement as the central weakness of the field, which means both sides agree on what the decisive experiment is.

Practice does not pick a winner here

You do, and your call is dated, kept, and yours to change. Follow the people you already trust and what they publish turns into a short daily practice.

Record where you land on this, and change your mind later without losing the history
The 7 practices that follow from this question, graded and linked to the source
Everything else these people publish, read and graded, in one place

Get started

Free to sign up and use. We do not ask for a card.