This free practice test for BCSP's Certified Safety Professional (CSP) exam holds 30 questions spread across the domains in roughly the proportions BCSP publishes. Once you choose, an explanation for each of the options opens below the item, so a miss shows which distinction you blurred. Nothing is gated and no account is needed.
Each lettered option unfolds to its own reasoning, kept apart from the other three.
Domain 1 · Safety principles · 25%Item 1 · 30 in the set
A main steam line has been isolated by two closed and locked block valves for a flange repair. What should be done BEFORE the flange is opened?
ARely on the two locked block valves
Why A is wrong. Two locked valves isolate the supply, but the pipe between them and the flange can still hold steam pressure.
BCrack one flange bolt and watch for leakage
Why B is wrong. Cracking a bolt and watching for leakage uses the worker's hands as the pressure test and releases any trapped steam right beside them.
CBleed and verify zero pressure and safe heat
Why C is right. Closed valves can still trap pressure and hot condensate between them, so the section is bled and zero pressure and a safe temperature are confirmed before the flange is opened.
DWait 30 minutes, then start the work
Why D is wrong. Waiting assumes the energy has dissipated; without a reading no one knows whether pressure or heat remains.
Domain 1 · Safety principles · 25%Item 2 · 30 in the set
Absence of voltage has been properly verified on a de-energized 13.8-kV overhead feeder that runs parallel to an energized line. What additional step is required before line workers contact the conductors?
APut on Category 2 arc-rated clothing
Why A is wrong. Arc-rated clothing protects against arc-flash heat, not against shock from voltage induced by the neighboring energized line.
BInstall temporary protective grounds
Why B is right. Grounds create an equipotential work zone against voltage induced by the parallel line and make protective devices trip if the line is accidentally re-energized.
CApply a second lock at the open switch
Why C is wrong. Another lock guards against switching errors at that point but does nothing about induced voltage from the parallel circuit.
DWait five minutes for the line to discharge
Why D is wrong. Waiting lets stored charge decay, but induction from the neighboring line continues for as long as that line is energized.
Domain 1 · Safety principles · 25%Item 3 · 30 in the set
A worker transfers 50-kg iron castings with an offset center of gravity, and each part must be rotated 90 degrees for assembly. Which lift assist is MOST appropriate?
AA vacuum tube lifter
Why A is wrong. A vacuum tube lifter needs a smooth sealing surface and handles offset loads poorly while they are turned.
BA chain hoist on a jib crane
Why B is wrong. A hoist lifts vertically but leaves the worker to steady and rotate a swinging, off-center load by hand.
CA spring-loaded tool balancer
Why C is wrong. Balancers support light hand tools, not 50-kg castings.
DA rigid-arm manipulator
Why D is right. A rigid-linkage manipulator resists the torque from an offset center of gravity and lets the worker rotate the part under control.
Domain 1 · Safety principles · 25%Item 4 · 30 in the set
An industrial robot has a maximum reach of 7 feet, and the workpiece it handles extends 3 feet beyond the gripper. To meet robot safeguarding requirements, where should the fixed perimeter fence be located?
AOutside the full 10-foot envelope
Why A is right. The hazard zone includes everything the robot can reach with its tooling and part: 7 ft of arm plus 3 ft of workpiece gives a 10-ft envelope, and the fence goes outside it.
BAt the 7-foot maximum reach line
Why B is wrong. Fencing at the arm's reach forgets the 3-ft workpiece, which can still strike someone standing at the fence.
CWithin the programmed 5-foot operating zone
Why C is wrong. Programmed limits can be exceeded through faults or reprogramming, and a fence inside the reach puts people within the swept area.
DAt the 18-inch clearance distance
Why D is wrong. A fixed 18-inch distance has no relation to how far this robot and its part can reach, so it would leave the swept area inside the fence.
Domain 1 · Safety principles · 25%Item 5 · 30 in the set
A process requires a relief flow of 0.022 m³/s. Three relief valves are available: Model X (certified 0.025 m³/s, backpressure correction factor 0.85), Model Y (0.028 m³/s, correction 0.90), and Model Z (0.030 m³/s, correction 0.95). Which valve provides adequate relief capacity with the minimum rated size?
AModel X
Why A is wrong. Its effective capacity is 0.025 × 0.85 = 0.02125 m³/s, below the required 0.022 m³/s once backpressure is accounted for.
BModel Y
Why B is right. Effective capacity is certified flow times the correction factor: 0.028 × 0.90 = 0.0252 m³/s, enough for 0.022 m³/s, and it is smaller than Model Z.
Model X derated=0.025 × 0.85 = 0.0213 m³/s
Model Y derated=0.028 × 0.90 = 0.0252 m³/s
Model Z derated=0.030 × 0.95 = 0.0285 m³/s
Smallest adequate valve=0.0252 ≥ 0.022 m³/s =Model Y
CModel Z
Why C is wrong. Model Z is adequate (0.030 × 0.95 = 0.0285 m³/s) but larger than needed, so it is not the minimum adequate size.
DBoth X and Y
Why D is wrong. Only Y is adequate; X falls short after the backpressure correction.
Domain 1 · Safety principles · 25%Item 6 · 30 in the set
Fleet telematics data reveals a specific pattern of frequent hard braking and rapid acceleration among delivery drivers. To address this behavior using leading indicators, which intervention strategy is most effective?
ADiscipline drivers with high braking scores
Why A is wrong. Punishment after the fact reacts to scores instead of changing driving as it happens, and tends to push drivers to hide or dispute the data.
BInstall speed governors on all vehicles
Why B is wrong. A governor caps top speed but does not stop harsh braking or rapid starts below that limit, and it ignores the individual pattern the data revealed.
CPost speed limit signs in the parking lot
Why C is wrong. Lot signage targets low-speed yard driving, while the data show braking and acceleration habits on delivery routes.
DInstall in-cab real-time coaching devices
Why D is right. Real-time in-cab feedback uses the same telematics signal to correct hard braking and rapid acceleration as they happen, turning the indicator into prevention.
Domain 1 · Safety principles · 25%Item 7 · 30 in the set
Operators on a packaging line are repeatedly bypassing a safety interlock by taping the sensor switch. Investigation reveals that the interlock trips frequently due to vibration, causing 15-minute restarts that prevent workers from meeting production quotas. What is the root cause of this unsafe behavior?
ASystem design error
Why A is right. Nuisance trips from vibration plus quotas that cannot absorb 15-minute restarts make bypassing the easiest way to get the job done, so the root cause lies in the equipment and work-system design.
BPoor supervision
Why B is wrong. Tighter supervision would only police the symptom; the faulty sensor and the quota conflict would remain.
COperator negligence
Why C is wrong. Blaming the operators stops the investigation at the person and leaves in place the conditions that make taping the sensor attractive.
DInsufficient training
Why D is wrong. The operators know the interlock and choose to defeat it; more training does not remove the vibration trips that drive the behavior.
Domain 2 · Program management · 25%Item 8 · 30 in the set
A procedure for charging a reactor has not been reviewed in five years but is still marked 'active'. Two process changes have been made under MOC since it was last revised. What should the safety manager do?
ARetrain operators on the procedure as it is written
Why A is wrong. Training on an outdated text spreads its errors to every operator.
BWithdraw it and rely on operator experience
Why B is wrong. Withdrawing the procedure and relying on experience leaves a hazardous task with no controlled instructions at all.
CArchive it as obsolete alongside the earlier versions
Why C is wrong. The task still exists, so archiving the only procedure without a revised replacement leaves operators without one.
DRevalidate it against the current process first
Why D is right. Two MOC changes since the last revision mean the text no longer reliably matches the process, so it is checked and revised against current conditions before anyone relies on it again.
Domain 2 · Program management · 25%Item 9 · 30 in the set
A safety interlock uses two independent sensors in parallel (either can detect the condition), followed in series by a logic solver. Sensor A has a reliability of 0.90, Sensor B 0.80 and the logic solver 0.99. What is the reliability of the interlock?
A0.7128
Why A is wrong. This treats all three components as one series chain (0.90 × 0.80 × 0.99), ignoring that either sensor alone can detect the condition.
B0.9702
Why B is right. The parallel sensors give 1 − (0.10 × 0.20) = 0.98, and in series with the logic solver R = 0.98 × 0.99 = 0.9702.
Both sensors fail=(1 − 0.90)(1 − 0.80) = 0.02
Sensor pair works=1 − 0.02 = 0.98
Interlock reliability=0.98 × 0.99 =0.9702
C0.9800
Why C is wrong. This is the reliability of the parallel sensor pair alone; the logic solver in series also has to work.
D0.9998
Why D is wrong. This treats the logic solver as a third parallel path, although its failure defeats the whole interlock.
Domain 2 · Program management · 25%Item 10 · 30 in the set
A chemical manufacturer's restructuring will reduce first-line supervisors by 30%. Which management of change step BEST addresses the main latent risk?
AMap each affected safety duty to a named owner
Why A is right. Removing supervisors can silently drop safety tasks such as permit sign-off and pre-job checks, so each duty is mapped to a named owner who is prepared before the change.
BSurvey employee morale six months after the change
Why B is wrong. A survey after the fact detects problems late and does not stop safety tasks from falling through the gap.
CNotify OSHA of the change in organizational structure
Why C is wrong. OSHA does not require notice of a reorganization, and notifying it would not assign anyone the lost tasks.
DRevise the corporate safety policy statement
Why D is wrong. A policy statement sets intent but does not decide who now signs permits or checks jobs.
Domain 2 · Program management · 25%Item 11 · 30 in the set
Two consecutive audits find that the annual ISO 45001 management review is run by the EHS manager alone, top management does not attend, and no decisions are recorded. Which corrective action MOST directly addresses the root cause?
AAdd a decision-minutes template
Why A is wrong. A decision-minutes template formats decisions that no one with authority is present to make.
BHave top management chair the review and decide
Why B is right. ISO 45001 clause 9.3 assigns the management review to top management, so the root cause is their absence; putting them in the chair, with recorded decisions, fixes it.
CTrain the EHS manager on the clause 9.3 requirements
Why C is wrong. The EHS manager is not the person the clause makes responsible, so better training leaves the review without anyone authorized to decide.
DAudit the review process and report quarterly
Why D is wrong. Quarterly audits and reports would only find the same gap again; they do not change who runs the review.
Domain 3 · Risk management · 15%Item 12 · 30 in the set
A company is adding a simple batch-mixing operation that closely resembles dozens it already runs safely. Which hazard analysis technique is usually MOST efficient for it?
AFault tree analysis
Why A is wrong. A fault tree analyzes one defined top event in depth rather than surveying all the hazards of a new operation.
BA full HAZOP study
Why B is wrong. A HAZOP's systematic node-by-node review suits complex or novel processes; here it would spend effort without adding insight.
CQuantitative risk assessment
Why C is wrong. Full quantification is kept for high-consequence or novel decisions; for a familiar, simple operation it costs far more than it adds.
DWhat-if/checklist analysis
Why D is right. For a simple process that closely matches many already run safely, a what-if/checklist review draws on that experience and finds the hazards with modest effort.
Domain 3 · Risk management · 15%Item 13 · 30 in the set
Two assessors rating the same crane-lift hazard on a risk matrix get very different severity ratings. The team leader sets a rule for the severity rating. Which consequence should the team normally rate?
AThe worst outcome that could credibly occur
Why A is right. Matrix severity is normally rated at the worst credible, meaning realistic, consequence, which keeps assessors consistent without ignoring serious outcomes.
BThe outcome of the most recent similar incident
Why B is wrong. The last similar incident is one sample and can be much milder or worse than what the hazard can realistically produce.
CThe average outcome of past incidents
Why C is wrong. Averaging past outcomes understates hazards that have so far been lucky.
DThe worst conceivable outcome, however unlikely
Why D is wrong. Rating at the worst conceivable outcome, however unlikely, pushes almost every hazard to catastrophic and destroys the ranking.
Domain 3 · Risk management · 15%Item 14 · 30 in the set
A company rates any potential fine over $10,000 as high risk. New regulations set minimum penalties of $50,000 for nearly every violation, so almost every compliance risk now lands in the top band and the ratings no longer separate them. Which element of the ISO 31000 process must be revisited FIRST?
AContext and risk criteria
Why A is right. The $10,000 line is a risk criterion, set with the organization's context; once nearly every violation crosses it, the criterion no longer separates risks and has to be recalibrated before anything is re-rated.
BRisk identification and analysis
Why B is wrong. Re-analyzing the risks against the same $10,000 line would put every compliance risk back in the top band; the line itself has to be reset first.
CSelection of risk treatment
Why C is wrong. Choosing treatments depends on a ranking that now puts almost every compliance risk in the same band, so it cannot come first.
DMonitoring and review
Why D is wrong. Monitoring and review is how the problem gets noticed, but the fix is made to the criteria themselves.
Domain 4 · Emergency management · 9%Item 15 · 30 in the set
A safety instrumented system logic solver damaged in a fire has been repaired. Before the system is returned to service, which verification is essential?
ASite acceptance and functional testing
Why A is right. After repair, the installed SIS has to be shown to perform its safety functions in place, which means site acceptance and proof or functional testing of the logic and field devices together.
BThe vendor's test and conformance certificates
Why B is wrong. Vendor certificates cover the product as shipped, not the repaired installation working with its field devices.
CThe factory acceptance test records
Why C is wrong. Factory acceptance tests were run before installation and before the fire, so their records say nothing about the repaired system.
DA visual inspection of the installation
Why D is wrong. A visual inspection of the cabinet and wiring cannot show that the logic, inputs and outputs perform their trips.
Domain 4 · Emergency management · 9%Item 16 · 30 in the set
A small ammonia release from a plant has drifted into a nearby neighborhood. What should the company's FIRST public message emphasize?
AProtective actions residents should take now
Why A is right. Residents exposed to a release need protective actions at once, such as sheltering in place or where to get help, alongside what is known so far.
BThe likely root cause found by investigators
Why B is wrong. The cause is not established in the first hours, and speculating risks corrections that erode trust.
CThe plant's strong prior safety record
Why C is wrong. Leading with the safety record sounds defensive and gives residents nothing they can act on.
DAssurance that the release posed no danger
Why D is wrong. Assurance of no danger before the facts are known can prove false and damages credibility for every later message.
Domain 4 · Emergency management · 9%Item 17 · 30 in the set
An on-site chemical fire was controlled quickly, but the emergency operations center did not know this and ordered an unnecessary community evacuation. Which plan element MOST likely failed?
AThe public warning system hardware
Why A is wrong. The warning system worked as ordered; the order itself rested on outdated information.
BThe ICS chain of command on scene
Why B is wrong. On-scene command controlled the fire, so the breakdown was not at the incident.
CInternal status notification
Why C is right. The EOC acted on stale information, so the failure lay in the protocol that passes status updates from the incident command post to the EOC.
DMutual aid agreements
Why D is wrong. Mutual aid brings in outside resources; the problem here was information that never reached the EOC.
Domain 5 · Environmental management · 6%Item 18 · 30 in the set
A former large quantity generator adopts just-in-time purchasing and chemical substitution. Its monthly hazardous waste generation over six months is 1,100, 950, 880, 1,050, 900 and 750 kg. What is its category and on-site accumulation limit for the most recent month?
ALQG; 90 days
Why A is wrong. Generator status is set month by month from that month's generation; earlier LQG months do not carry over to a 750-kg month.
BSQG; 90 days
Why B is wrong. The category is right, but 90 days is the large quantity generator limit.
CSQG; 180 days
Why C is right. 750 kg is between 100 and 1,000 kg, so the facility is a small quantity generator for that month, with up to 180 days of on-site accumulation (270 days if the receiving facility is more than 200 miles away).
DVSQG; no time limit
Why D is wrong. VSQG status needs 100 kg or less in the month, and 750 kg is well above that.
Domain 5 · Environmental management · 6%Item 19 · 30 in the set
A municipal-scale treatment plant's permit adds a new total nitrogen limit. Which process is MOST commonly used to meet it?
ANitrification alone in the aeration basin
Why A is wrong. Nitrification converts ammonia to nitrate, so the nitrogen changes form without leaving the water and total nitrogen barely drops.
BNitrification followed by denitrification
Why B is right. Total nitrogen removal takes both steps: nitrification oxidizes ammonia to nitrate aerobically, then denitrification reduces the nitrate to nitrogen gas under anoxic conditions.
CLonger settling time in the clarifiers
Why C is wrong. Settling removes solids; most of the nitrogen is dissolved and passes through the clarifier.
DChemical precipitation with alum
Why D is wrong. Alum precipitation is a phosphorus removal method and does little for nitrogen.
Domain 6 · Occupational health · 10%Item 20 · 30 in the setBeryllium
Medical surveillance at a beryllium-copper machining shop finds a worker sensitized to beryllium, although every personal air sample for that worker is below the OSHA action level of 0.1 µg/m³. Which exposure route is the MOST likely contributor?
AIngestion of dust while eating in the work area
Why A is wrong. Beryllium is poorly absorbed from the gut, and ingestion is not a recognized sensitization route.
BRemoving the respirator briefly during tool changes
Why B is wrong. Personal samples are collected outside the respirator, so brief removal is already reflected in the low air results.
CEye contact with dust and chips during grinding
Why C is wrong. Dust and chips in the eye cause local irritation; the documented route to sensitization besides inhalation is skin contact.
DSkin contact with contaminated dust and fluids
Why D is right. Beryllium sensitization can develop through the skin as well as by inhalation, which is why the OSHA standard requires gloves, protective clothing and hygiene controls along with air controls.
Domain 6 · Occupational health · 10%Item 21 · 30 in the set
Office workers seated beside a large west-facing glass wall complain of feeling hot every afternoon, although the thermostat reads 72°F (22°C) and relative humidity is 45%. Which thermal comfort factor BEST explains the complaints?
AHumidity outside the comfort range
Why A is wrong. The 45% relative humidity sits inside the usual comfort band, so humidity does not explain the complaints.
BInadequate outdoor air raising CO2 levels
Why B is wrong. Low ventilation causes stuffiness and air-quality complaints rather than a heat sensation tied to the afternoon sun.
CElevated mean radiant temperature from solar gain
Why C is right. Afternoon sun through a large glass wall raises the radiant heat load on people next to it even when the air temperature at the thermostat is normal.
DExcessive air velocity at the workstations
Why D is wrong. High air speed cools people and brings draft complaints, not a feeling of heat.
Domain 6 · Occupational health · 10%Item 22 · 30 in the set
An analysis of a lifting job shows that no single lift comes close to spinal tissue tolerance, yet workers who have held the job for years develop disc injuries. Which concept BEST explains this pattern?
AA single lift that exceeded tolerance
Why A is wrong. The analysis already shows no single lift comes near tissue tolerance, so a one-time overload does not fit.
BCumulative damage from repeated submaximal loads
Why B is right. Repeated submaximal loads well below single-lift tolerance cause cumulative microdamage and lower tissue tolerance over time, much like fatigue failure in metal under cyclic stress.
CWhole-body fatigue from high metabolic workload
Why C is wrong. Metabolic fatigue concerns energy expenditure, not mechanical wear of spinal tissue.
DPsychosocial stress lowering pain thresholds
Why D is wrong. Stress affects how symptoms are reported but does not produce the structural disc damage described.
Domain 1 · Safety principles · 25%Item 23 · 30 in the set
A worker wears an arc-rated shirt rated 5 cal/cm² under an arc-rated coverall rated 4 cal/cm². The task requires arc-rated clothing of at least 8 cal/cm². Does this combination meet the requirement?
AYes, because 5 + 4 = 9 cal/cm²
Why A is wrong. Arc ratings of separate garments cannot be added, because air gaps and materials interact; the combination needs its own tested rating.
BYes, if the outer layer is the higher-rated one
Why B is wrong. Layer order does not create a rating; only a test of the combination as worn does.
CNo, because layering arc-rated clothing is prohibited
Why C is wrong. Layering arc-rated clothing is permitted and common in cold weather; the issue is how the system's rating is established.
DNo, unless the system test rating reaches 8
Why D is right. Layered arc-rated clothing is rated as a system by testing the combination as worn, so it meets an 8 cal/cm² requirement only if that system test rating reaches 8.
Domain 2 · Program management · 25%Item 24 · 30 in the set
A safety manager updates a critical lockout/tagout procedure by overwriting the old file with the new text and changing the document date. From a document control perspective, what critical element is missing?
AA revision history with the approval record
Why A is right. Document control needs a record of what changed, why and who approved it; overwriting the file destroyed both the prior version and that trail.
BA scheduled date for the next periodic review
Why B is wrong. A review date keeps procedures current, but it does not show what changed in this revision or who approved it.
CPosting to the intranet library
Why C is wrong. Posting the file to the library makes the procedure available but leaves no trace of the earlier version or of the approval.
DA signature block for the author on each page
Why D is wrong. Author signatures identify who wrote the text, not what was changed or that the change was authorized.
Domain 2 · Program management · 25%Item 25 · 30 in the set
At an audit's closing meeting, the plant manager disputes a nonconformity the audit team has reported. According to ISO 19011, what should the team do?
AWithdraw the disputed finding
Why A is wrong. Dropping a finding backed by evidence to keep goodwill gives up the evidence-based basis of the audit.
BPause the meeting and wait for agreement
Why B is wrong. Pausing until the manager agrees hands the auditee a veto over the findings and undermines the audit's objectivity.
CDiscuss and record any unresolved opinion
Why C is right. ISO 19011 says diverging opinions on findings should be discussed and resolved if possible, and any that remain unresolved are recorded.
DRefer the finding to the certification body
Why D is wrong. The guideline handles a disagreement inside the audit by discussion and a record; it has no step for handing the finding to an outside body.
Domain 2 · Program management · 25%Item 26 · 30 in the setPSI
An engineering team proposes upgrading a process pump to a larger model with a bigger motor and different flow characteristics. In addition to the operating manual, which set of process safety information (PSI) must be updated before the Management of Change (MOC) can be closed?
ATraining records and shift logs
Why A is wrong. Training records matter under MOC, but they are not process safety information and leave the drawings and relief design untouched.
BP&IDs, electrical data, and relief design
Why B is right. A larger pump changes the drawings, the electrical information and the flow the relief system must handle, so the P&IDs, electrical data and relief design basis all need updating before closure.
CThe pump and motor data sheets only
Why C is wrong. The pump and motor data sheets are part of the update, but this set leaves out the P&IDs and the relief design that the new flow affects.
DSDSs, chemical inventory, and container labels
Why D is wrong. The chemicals have not changed, so the SDSs, chemical inventory and container labels stay valid; the change is to equipment.
Domain 2 · Program management · 25%Item 27 · 30 in the setMetrics
A safety dashboard displays the following four metrics. Which one is considered a "vanity metric" because it provides a false sense of success without actionable context or normalization?
AClosure rate of corrective actions
Why A is wrong. A closure rate is a ratio tied to action, showing whether identified problems are being fixed.
BNear-miss reports per 200,000 hours
Why B is wrong. Near-miss reports scaled per 200,000 hours of exposure can be compared across periods and sites, so they carry the context a raw count lacks.
CTotal count of observations submitted
Why C is right. A raw count has no denominator or quality measure, so it can climb simply because more cards are turned in, giving a look of progress without telling anyone what to change.
DRate of high-risk findings per audit
Why D is wrong. Normalizing by audits makes the figure comparable and points directly at where serious hazards are being found.
Domain 3 · Risk management · 15%Item 28 · 30 in the set
A property insurer estimates the Maximum Foreseeable Loss (MFL) for a sprinklered warehouse. Which assumption does the MFL estimate make?
AAll fire protection systems operate as designed
Why A is wrong. An estimate with all fire protection operating as designed is the normal loss expectancy, the smallest of the insurer's loss figures.
BOne critical system is impaired and the rest work
Why B is wrong. One impaired critical system with the rest working is the basis of the probable maximum loss, which sits between the normal loss expectancy and the MFL.
CFixed protection fails and fire walls stop it
Why C is right. MFL assumes the automatic protection is impaired, so the fire burns until passive barriers such as fire walls or space separation stop it; it is the largest loss expected at that location.
DOnly the building is lost, not its contents
Why D is wrong. Loss estimates cover building and contents together; the MFL is defined by which protection is assumed to fail, not by leaving out the contents.
A safety manager plans tablet-based microlearning for production workers. Wi-Fi coverage and device budget are already confirmed. Before buying the tablets, which factor should be assessed to ensure the program works?
AStock of replacement parts
Why A is wrong. Spare parts keep devices running but do nothing for workers who cannot navigate them.
BWorkers' preferred tablet brand
Why B is wrong. Workers' preferred tablet brand is a matter of taste; it says nothing about whether they can operate the tablets and the content.
CSoftware license costs
Why C is wrong. License costs are a budgeting detail; they do not decide whether workers can learn from the tablets.
DWorkforce digital literacy
Why D is right. With coverage and budget settled, the remaining risk is the workforce's digital literacy: if production workers cannot operate the tablets and content, delivery fails regardless of the hardware.
When setting pass/fail criteria for the practical assessment of a high-risk task such as lockout/tagout, which standard is MOST appropriate?
AA 70% score on the task checklist
Why A is wrong. A percentage cut-off lets someone pass while skipping a step such as verifying zero energy.
BZero errors on all critical steps
Why B is right. For lockout, a single missed critical step such as verifying zero energy can be fatal, so the passing standard is flawless performance of those steps.
CTask completion within ten minutes
Why C is wrong. A ten-minute completion limit measures speed, which says nothing about whether each energy source was isolated and verified.
DA pass on the written knowledge test
Why D is wrong. A written test shows what the person knows, not that they perform the task correctly.
BCSP weights its domains unevenly, and the set leans the same way. Most questions come from Advanced Application of Safety Principles and Program Management, which hold 25% of the exam apiece, while Environmental Management and Training supply two items each. Each item carries a domain chip, so a cluster of misses sends you to the right part of the study guide.
Most items are short workplace scenarios that ask for the best next step or the most likely cause, the management angle the blueprint stresses. A handful need arithmetic.
A first pass without notes gives the most honest picture. Second time through, read the reasoning under every option of the items you missed, including the choices you never considered: each explains why that choice fails for this particular scenario, which is often a near neighbor of the right answer.
Track misses by domain chip rather than by total. Two misses in a domain worth 6% weigh less on the full exam than two in one worth 15%, and the chip tally tells you where an hour of review buys the most. The total itself cannot be converted into a BCSP result, because BCSP keeps its cut score private and scores a full form of 200 items as one sum.
On the full exam you meet 200 items, each with 4 choices and one key, presented singly on screen. When an item needs a formula, BCSP prints it in the item, unless the formula itself is what is being tested. For arithmetic you get an on-screen copy of the TI-30XS scientific calculator; you may not bring your own. The exam facts page covers timing, breaks and scoring.
The items here are practice questions, not BCSP questions, and none is copied from an exam form. BCSP guards its item bank, and candidates agree not to reveal exam content; BCSP takes legal action over shared items. How the set was prepared and checked is summarized in the table below and on the about page.
Fact
Detail
Source
About our questions
Exambrik’s practice questions are not BCSP exam items. They were adapted from existing exam-prep question banks. Before publication each one was worked without the answer key, screened for any other option that could also hold up, checked against the current CSP blueprint and the regulation or consensus standard as it reads now, and paired with an explanation for each choice. BCSP’s self-assessments and examCORE questions belong to BCSP and are not reproduced here.
BCSP’s paid CSP examCORE course (online, self-paced, pre- and post-tests, over 1,000 questions) and a half-length online self-assessment. BCSP keeps the examCORE authors apart from its exam-development panels and says finishing the course is no promise of passing.
About this Certified Safety Professional practice set
How does this set differ from BCSP's own self-assessment?
The board's self-assessment is paid, half the length of the full exam, and sold alone or inside its $600 bundle with six months of access. This set is free, much shorter, and explains each option right after you answer. Neither one is a copy of the live exam.