SB 326 & SB 721 sample size calculator
Size the sample, build the randomized element list the statute requires, and draft the report’s page one. Free, no sign-up, and nothing about your building leaves this tab.
If you take one number away: on a 200-element complex, SB 721’s 15% is 30elements and SB 326’s “statistically significant” sample is 132. They are different tests and they are not close.
Civil Code § 5551. A random, statistically significant sample of the association’s exterior elevated elements — 95% confidence, ±5% margin of error. Licensed structural or civil engineers and architects only. First inspection was due 1 January 2025, then every 9 years. Both cover elements more than 6 ft above ground and substantially supported by wood, plus their waterproofing.
Element inventory
Prefilled with an example 200-element complex so there is a number to look at. Type over it — nothing here is sent anywhere.
How the types are drawn changes the SB 721 answer, because its 15% is per type — that division is your professional judgement, not ours. Rows with the same name are merged, since two rows of 10 “Balconies” would otherwise each round to 2 and return 4 of 20 where the statute asks for 3.
Sampling parameters
z = 1.9600, assumed population proportion 0.5— the value that maximises the required sample, and the only defensible one before any element has been looked at. Finite-population correction applied to N = 200.
- 132elements to inspect of 200
- 66.0%of the population
- 4element types in scope
This is 4.4× the 15% figure. SB 721’s 15% on the same 200 elements would be 30. SB 326 asks for 132, because 95%/±5% on a population this small is 66.0%of it — not 15%. Pages that quote 15% for both statutes are describing the wrong one. Unconstrained, the same parameters would need 385 elements; the finite-population correction is what brings it down to 132.
| Element type | In building | Inspect | Share |
|---|---|---|---|
| Balconies | 120 | 79 | 65.8% |
| Stairways | 48 | 32 | 66.7% |
| Elevated walkways | 26 | 17 | 65.4% |
| Porches | 6 | 4 | 66.7% |
| Total | 200 | 132 | 66.0% |
Randomized element list
SB 326 does not only ask for a random sample. It requires a randomized list of every element’s location, given to the association, with the next inspection resuming at the next element on that list— nine years later, possibly by another firm. So the order has to be reproducible from the record, which is why it comes from a seed you choose and put in the report rather than from a shuffle nobody can re-derive.
Enter a seed to generate the list. Any short phrase works; the same phrase always produces the same order, which is the whole point.
Report page one
§ 5551 specifies what page one of the report must state. These are those facts in the statute’s own order, for pasting into your own template. Fill in what the calculator cannot know.
Inspection date: — (to be completed) Total units in the building: — (to be completed) Units with exterior elevated elements: — (to be completed) Total exterior elevated elements: 200 Total elements inspected: 132 (66.0% of the population) Elements posing an immediate threat to safety: — (to be completed) Sample certification The inspected elements were drawn from a randomized list of all 200 exterior elevated elements (seed "(not yet chosen)", reproducible), sized for 95.0% confidence at a ±5.0% margin of error, assuming a population proportion of 0.5 (the value that maximises the required sample), with a finite-population correction applied to N = 200. Required sample: 132 elements.
This sizes a sample. It is not a professional opinion.It does not determine what to inspect, what condition anything is in, or whether a report satisfies a statute — the licensee signing the report decides all of that, and remains responsible for it. Statutory text and deadlines change; verify § 5551 and § 17973 against the current code before relying on any figure here.
The two statutes, side by side
| SB 326 — Civil Code § 5551 | SB 721 — Health & Safety § 17973 | |
|---|---|---|
| Applies to | Condominium and other HOA-governed buildings | Buildings with 3+ multifamily dwelling units, excluding condos |
| Sample | Random and statistically significant — 95.0%confidence, ±5.0% margin of error, on the population of elements | At least 15.0% of each type of element |
| Who may inspect | Licensed structural or civil engineers, and architects | Architects; civil or structural engineers; certified building inspectors; or A, B or C-5 contractors with 5+ years of multistory wood-frame experience |
| First deadline | 1 January 2025 — passed | 1 January 2026 — passed |
| Then every | 9 years | 6 years |
| Element list | Randomized list of everyelement’s location, given to the association; the next cycle resumes at the next element on it | Not specified |
| In scope | Balconies, decks, porches, stairways, walkways and their railings, more than 6 ft above ground and substantially supported by wood — plus the associated waterproofing. | |
Statutory text checked 2026-09-13. Deadlines and thresholds get amended; verify against the current code before you rely on a row.
Why 15% is the wrong number for SB 326
The sample size for a proportion at a given confidence and margin is the familiar one:
n₀ = z² · p(1 − p) / e² z = 1.9600 (two-sided, 95.0% confidence) p = 0.5 (maximises p(1 − p), so it maximises n) e = 0.05 (±5.0%) n₀ = 385 elements
That 385 is where most published answers stop, and it is why some of them say every building needs about 385 elements inspected. It is the sample size for an infinite population. A condominium complex is not one, and there is a correction for that:
n = n₀ / (1 + (n₀ − 1) / N)
Apply it and the answer depends on the building, which is the whole point. The fraction falls as the building grows — a large portfolio building gets off comparatively lightly, a small association does not:
| Elements (N) | SB 326 sample | Share | SB 721’s 15% would be |
|---|---|---|---|
| 50 | 45 | 90.0% | 8 |
| 100 | 80 | 80.0% | 15 |
| 200 | 132 | 66.0% | 30 |
| 500 | 218 | 43.6% | 75 |
| 1000 | 278 | 27.8% | 150 |
Below about 385 elements the correction is doing most of the work; above it, the sample flattens toward 385 no matter how large the property gets. A 20-element association needs all 20 — at that size there is no sample that supports the claim, only a census.
Questions
Is a p of 0.5 not overly conservative?
It is the most conservative choice and that is why it is the default. A smaller p means a smaller sample, but justifying one means asserting in advance what fraction of the elements is sound — which is the finding the inspection exists to produce. At a second cycle you have a measured prior from the first, and using it is defensible. The calculator lets you, and marks the result as departing from the statutory parameters when you loosen confidence or margin.
Does SB 326 require the sample to be stratified by element type?
No. It sizes on the population of exterior elevated elements, unlike SB 721, which is explicitly per type. The calculator allocates the total across types in proportion anyway, because a sample of 132 drawn entirely from balconies is statistically significant about balconies and silent about the stairways, and that is not a report anyone stamps. Stratifying is a strengthening of the statute, not a requirement of it.
Why does the randomized list need a seed?
Because § 5551 requires the next inspection to resume at the next element on the list — nine years later, quite possibly by a different firm. A shuffle nobody can re-derive satisfies “random” and fails “resume”. A seed printed in the report regenerates the exact same order forever, with no database and no dependence on this page still existing. Record the seed and the cursor position; that is the entire handoff.
Is this legal advice, or an engineering opinion?
Neither. It sizes a sample and orders a list. It does not decide what to inspect, what condition anything is in, or whether a report satisfies a statute. The licensee who signs the report decides all of that and remains responsible for it.
Are you going to sell my details?
The calculator posts nothing. No email is asked for, there is no form on this page, and your element counts are state in your own browser tab. We are not a lead broker and never have been — the business is measured scopes, priced per job.
The part that actually costs you the day
Sizing the sample takes a minute. Producing the element inventory, the measured per-element baseline and the photographs is the week.
Nuvea turns a phone scan of a building into measured dimensions in feet and inches, an itemized takeoff, and a staged scope you can price. It was built for remodelers scoping work; the same measurement, per element, is the baseline these statutes ask for. That fit is a claim we are currently testing with inspection firms rather than one we are advertising as finished — if you do this work and want to be one of them, the honest offer is that your first building is free and we will tell you plainly what does not work yet.
Nuvea is not a licensed inspector and does not issue reports. It measures and drafts; you inspect, and you sign.