CSP-11 CSP-11 D2.14 · Domain 2 · Program Management · 25%
CI / Probability
A packaging supervisor cites zero recordable cases in 20,000 worker-hours to postpone guarding a conveyor nip. Decide what the zero estimate establishes, what remains uncertain, and whether either permits exposed machinery.
Updated . Independent CSP-11 practice. Not BCSP-endorsed. No pass guarantee.
Decision rule
A confidence interval describes a sample statistic, not a guaranteed individual.
CI width falls when n rises or variability falls. Probability on a stem is not a feeling.
In 20 seconds
- A confidence interval describes a sample statistic, not a guaranteed individual.
- Probability quantifies likelihood; independence assumptions must be justified.
- Confidence intervals express estimate precision — not magic certainty.
- Do not confuse confidence level with probability a single true value is inside once realized.
How to apply CI / Probability
THEORY: Sampling variation. A 95% CI is a method property, not ‘95% sure this worker is safe.’ Independent events multiply; mutually exclusive events add. CALC: If a stem gives mean, SD, n, use the given interval. Teaching P(A and B)=P(A)P(B) only when independence is stated. TI-30XS: use given numbers only. FIELD: IH wants to declare a process ‘in control’ from n=3. Do not over-claim; collect a designed sample. ASSESS: Treating a CI as a PEL is the trap.
Worked case
Wrong first move: Cite the CI as proof of compliance. Correct move: State what the interval estimates and what n cannot support.
Why the rule holds
Related: 36, 62, 92.
Same rule, different setting
A pad had zero recordables in 20,000 hours and claims we are 95 percent sure TRIR is 0 forever. You are the CSP. The question is what a confidence interval and a probability are allowed to say.
A job had zero recordables in 20,000 hours and claims we are 95 percent sure TRIR is 0 forever. You are the CSP. The question is what a confidence interval and a probability are allowed to say.
Picture / model
Define a failure event and its exposure basis, choose the correct AND/OR identity, challenge independence, and distinguish estimated probability, event rate, and confidence-interval coverage before making a safety decision.
Exam traps
Original SafePath contrasts for CI / Probability. These are study traps, not recalled BCSP items.
| Trap | Correct read |
|---|---|
| Independent AND | Multiply (if the stem independence is real) |
| Mutually exclusive OR | Add |
| 95 percent CI | Method covers the parameter 95 percent of the time in the long run |
| P=0 from 0 events | Point estimate only — CI still wide |
Field move
P(A and B) = 0.0006 while independence holds.
Scenarios
Scenario 1 · Fundamental · CSP-11 D2.14
You are the CSP reviewing a live decision at a Gulf Coast chemical terminal. Operations proposes a convenience path. If A and B are independent with P(A)=0.02 and P(B)=0.03, what is P(A and B)?
Best action: 0.0006 (multiply).
Why: Independence gives P(A and B) = 0.02 × 0.03 = 0.0006. Adding the probabilities is the mutually exclusive OR rule, not this AND.
Scenario 2 · Applied Scenario · CSP-11 D2.14
You are the CSP reviewing a live decision at a refinery coker outage. Operations proposes a convenience path. If A and B are mutually exclusive with P(A)=0.02 and P(B)=0.03, what is P(A or B)?
Best action: 0.05 (add).
Why: Mutually exclusive events have no overlap, so P(A or B) = 0.02 + 0.03 = 0.05. The product 0.0006 would be an independent AND, and the general OR formula subtracts an overlap that the stem says is zero.
Scenario 3 · Applied Scenario · CSP-11 D2.14
0 recordables in 20,000 hours. Claim: 95 percent CI proves TRIR is 0 forever; skip the guard. Best call?
Best action: Reject: point estimate 0 with small EH is weak; a CI is not a 95 percent promise of the future; specify the still-open guard.
Why: Evaluate the best response. Zero recordables in 20,000 hours is a small exposure. A confidence interval on that point is not a 95 percent promise that TRIR stays zero, and the nip is still unguarded. Treat zero observed cases as a zero upper bound on risk. — Treating zero observed cases as a zero u…
If the stem says this
| Stem signal | Best direction |
|---|---|
| events are not independent | do not multiply as if independent |
| severity is catastrophic despite low probability | still prioritize hierarchy-smart controls |
| a CI is wider after a smaller sample | expect less precision — do not invent certainty |
| two answers work | higher hierarchy / system / design |
| the stem asks a simple independent AND | multiply the probabilities |
Check before you leave
- ✓ Probability quantifies likelihood; independence assumptions must be justified.
- ✓ Confidence intervals express estimate precision — not magic certainty.
- ✓ Do not confuse confidence level with probability a single true value is inside once realized.
- ✓ Rare high-severity events still need controls despite low frequency.
- ✓ Bayes-style updating is conceptual — read the stem carefully.
- ✓ AND of independent events multiplies; OR adds with care for overlap.
Blueprint mapping
CI / Probability is the SafePath study class for CSP-11 D2.14 inside CSP-11 Domain 2, Program Management (25%). A high-weight management domain focused on measuring performance, investigating incidents, managing change, building systems, leading people, controlling records and resources, and interpreting data.
SafePath uses the publicly posted CSP-11 outline as a study map. It does not reproduce official exam content, and it is not affiliated with or endorsed by BCSP or Pearson VUE.
Practice / flashcards / formulas
More classes in Program Management
Nearby lessons mapped to the same CSP-11 domain.
CSP-11 D2.14
Other classes on this task
Guides that use CI / Probability
- Oil and gas safety study guideProcess safety, change, and investigations
- CSP application and exam guideEligibility, scheduling, and exam method
- CSP-11 exam prep guideDomain weights, sample items, and a study sequence
- CSP-11 blueprint and weightsSeven published domains and the class map
- Seven-domain study directoryScope, tasks, and links into each lesson
CI / Probability study questions
- What decision does the CI / Probability class train?
- A confidence interval describes a sample statistic, not a guaranteed individual. This is independent SafePath CSP practice for CSP-11 D2.14. It is not an official BCSP item and it is not a pass guarantee.
- What exam trap shows up in CI / Probability?
- A common miss treats this as sufficient: Independent AND The stronger read is: Multiply (if the stem independence is real)
- How does CI / Probability map to the CSP-11 blueprint?
- SafePath maps this class to CSP-11 D2.14 in Domain 2, Program Management (25% of the published blueprint). Use the public outline for study planning. SafePath is not affiliated with or endorsed by BCSP.
Verify
Verify: Confirm current certification requirements with BCSP and technical requirements with the cited authoritative standard.