Sex as a Biological Variable in NIH Grants: How to Write the SABV Plan Reviewers Accept

What the SABV Policy Actually Asks For

NIH expects applicants to consider sex as a biological variable in the design, analysis, and reporting of studies involving vertebrate animals and human subjects. The expectation took effect for applications in 2016 and has been folded into how reviewers think about rigor ever since. Read the wording closely, because people tend to remember a stricter rule than the one that exists.

The policy doesn't say you must study males and females in every experiment. It says you must factor sex into the design and be able to explain what you did. Strong justification for studying one sex is allowed. Silence is what gets you in trouble. A plan that never mentions sex forces the reviewer to guess whether you forgot or decided not to bother, and reviewers rarely guess in your favor.

One caveat. Exact application instructions change with each forms release, so check the current instructions and your institute's notices before you finalize wording. I'm describing the stable core of the policy here, not a specific form field.

Where SABV Lives in the Application

There's no standalone SABV page for most mechanisms. It shows up in three places. First, the Research Strategy, usually in the rigor discussion within the Approach. Second, the Vertebrate Animals section, where you describe animal numbers and justify the species and sexes you've chosen. Third, for human studies, the Human Subjects and Clinical Trials information, where sex or gender enrollment planning sits alongside the inclusion plan.

Under the simplified review framework, rigor feeds the Approach factor, and the Approach factor is where most of the score pressure lands. So an SABV weakness doesn't sit in a corner of the summary statement. It tends to show up in a reviewer's Approach bullets, next to the criticisms that move your impact score. If you want a refresher on how the pieces connect, our simplified review framework guide walks through the factors.

The practical move is to say the same thing consistently in all three places. Reviewers read the Vertebrate Animals section and the Approach separately, and a mismatch between them (equal numbers of males and females described in one, only males in the other) is exactly the kind of inconsistency that gets flagged.

Writing an SABV Plan That Isn't a Placeholder

The weak version is familiar: "Sex will be considered as a biological variable." It's true, it's short, and it tells the reviewer nothing. Compare it with a plan that commits to specifics.

That last bullet is the one applicants skip, and it's the one reviewers like most. It shows you've thought about the result, not only the checkbox. You don't need to overpromise. "If we observe a sex-by-treatment interaction, we will report effect sizes by sex and flag it as a hypothesis for follow-up" is honest and sufficient.

Justifying a Single-Sex Design

Single-sex studies are acceptable when there's a real reason. Real reasons include a disease that occurs in one sex, a model that only exists or is validated in one sex, a hormone-dependent mechanism where the question is specific to one sex, or published data showing the phenotype is absent in the other sex. Say which one applies and cite the evidence.

Weak reasons are the ones that quietly sink the argument. "Females introduce variability from the estrous cycle" is the classic. Reviewers have heard it many times, and there's published work suggesting the added variability in female rodents is often not larger than in males, though it depends on the trait. "We only have the male line" is honest but reads as convenience. "Cost" on its own tends to land badly unless you've shown that adding the other sex would break the budget.

If your honest reason is practical, consider a middle path. Run the full study in one sex and include a smaller, pre-specified confirmation cohort in the other for the key readout. It's not perfect, but it demonstrates intent, and a reviewer can see exactly what you've committed to. I can't promise a given panel will accept it. Some will, and some will still want both sexes throughout.

Power, Sample Size, and the Budget Conversation

The objection I hear most from lab heads is that doubling the experiment to include both sexes doubles the cost. Usually it doesn't. If you're adding sex as a factor, you split your planned group size across the two sexes while keeping the total in the same neighborhood, and you power the primary comparison on the pooled effect. What you give up is the ability to detect a modest sex-specific difference with confidence, and you should say so plainly.

Write that tradeoff into the application. Something like: "The study is powered to detect the primary treatment effect across sexes; it is not powered to detect small sex-by-treatment interactions, which we will treat as exploratory." That sentence protects you. A reviewer who wants interaction-level power can see you understood the limitation, and it's much harder to call a transparent design a rigor failure. If the interaction itself is a central hypothesis, then you do need to power for it, and the sample size justification has to reflect that.

For budget, keep the language consistent with your budget justification. If animal numbers or per-diem costs scale with the design, the justification should say so, so the scientific plan and the dollar figures tell the same story.

Mistakes That Draw Rigor Critiques

Before you submit, run a final read of the Approach with one question: could a reviewer tell, for every aim, which sexes are included and why? If the answer takes more than ten seconds to find, tighten it. Our final-week checklist is a good place to add that check.

Frequently Asked Questions