Thirty six recreational athletes from Stanford, Rutgers, Vanderbilt and the University of Reading ate from their university dining halls for two randomized four week blocks, one omnivorous and one plant based, serving as their own controls.
Runners' 12 minute timed run differed by 36.2 m between diets (95% CI -130.3 to 57.9 m, p=0.43), and composite machine strength differed by 4.0% (95% CI -8.5 to 0.6%, p=0.08). Neither reached significance.
Protein intake was substantially lower on the plant based block, 71 g per day against 103 g per day, while remaining within recommendations. The plant based phase was higher in carbohydrate, fiber, iron and calcium.
The numbers
| 12 minute timed run, mean difference omnivore minus plant based | -36.2 m, 95% CI -130.3 to 57.9 m, p=0.43 |
| Composite machine strength difference | -4.0%, 95% CI -8.5 to 0.6%, p=0.08 |
| Protein intake | 71 g/day plant based vs 103 g/day omnivore |
| Plant based block nutrient profile | higher in carbohydrate, fiber, iron and calcium |
What this does not show
- That the diets are equivalent. This is a null result at n=36, and the confidence intervals are wide enough to contain effects that would matter to a competitive athlete. Absence of a detected difference is not equivalence.
- That strength was unaffected. The strength result sits at p=0.08 with a CI from -8.5% to +0.6%, very nearly all on one side of zero. The honest reading is a possible small strength decrement this study was not powered to resolve, not no difference.
- That intakes were accurately recorded. The authors flag that reported calorie totals were unusually low, which means dietary intake data were probably under recorded. That weakens the protein comparison specifically.
- Anything about elite athletes or longer timescales. Four weeks in recreational undergraduates training three to four times weekly. Muscle adaptation, iron status and micronutrient stores operate on longer timescales, and the authors say so.
- That a 12 minute run is a marathon. Timed run and three repetition maximum are crude proxies for endurance and strength performance. There was also no blinding and no described washout, both unavoidable in a dining hall study and both real limitations.
Caveats worth holding
- Date trap: the paper was published 20 February 2026; the Rutgers press release of 4 August 2026 is a news hook, not a new study. Cite the February date.
- A null result at n=36 with wide confidence intervals is not evidence the diets are the same.
- The strength comparison at p=0.08 is consistent with a small decrement on the plant based diet that the study could not resolve.
- Self reported dietary intake appears under recorded, per the authors, weakening the protein comparison.
- Four weeks is short relative to muscle adaptation, iron status and micronutrient stores.
- Recreational undergraduates; says nothing about elite or high volume endurance athletes.
- No blinding, no described washout, and crude performance proxies.