Certified Information Systems Auditor (CISA) — Free Practice Questions
10 free sample questions from a bank of 2264, with the correct answers and explanations. No signup required — start practising right now.
1Which of the following should be of GREATEST concern to an IS auditor reviewing an organization's business continuity plan (BCP)?
The BCP has not been tested since it was first issued.
The BCP is not version-controlled.
The BCP's contact information needs to be updated.
The BCP has not been approved by senior management.
Answer: A
The short version
A — An untested BCP is the greatest audit concern. A plan that never faced a drill may fail when needed. Testing proves recoverability, other gaps are secondary.
Key concepts in this question
Business continuity plan: documented procedures to sustain critical processes through disruption and restore normal service.
Testing value: exercises reveal missing steps, stale contacts, resource shortfalls and timing errors before a real event.
Version control: tracks revisions for accountability, but an untested current version still carries higher risk.
Contact hygiene: outdated phones slow notification yet can be worked around during response.
Approval signal: senior sign-off shows support, but approval without validation does not prove the plan works.
Why A is correct
A is correct because ISACA ranks an untested plan as the top failure mode, since unproven recovery steps give false assurance. Standards require periodic tests, from tabletop to full interruption, to confirm roles, dependencies and timing. A BCP untouched since issuance likely reflects changed systems and staff, so recovery time goals are fiction. Versioning, contacts and approval matter, yet none substitutes for evidence that the plan actually restores operations.
Why the others are wrong
B. lacks version control, which hurts traceability but does not prove inability to recover like missing tests do.
C. needs contact updates, an administrative fix with lower impact than unvalidated procedures.
D. lacks senior approval, a governance gap that is serious yet less critical than unknown operability.
CISA exam tip
Greatest concern equals operability risk. Choose untested over unapproved, uncontrolled or outdated contacts every time.
2Which of the following is the BEST way to ensure that an application is performing according to its specifications?
Pilot testing
System testing
Integration testing
Unit testing
Answer: B
The short version
B — System testing. Conformance to specifications is proven by testing the complete, integrated application against those specifications.
Key concepts in this question
Test levels: Unit (internal logic of one module), integration (interfaces between modules), system (the whole application against stated requirements), acceptance/pilot (limited user trial before rollout).
Specifications: The functional and performance requirements written at requirements analysis; a control is only meaningful if the test asserts against them.
Separation of concern: Only system-level testing pulls in hardware, OS, interfaces and data, so it is the only level that can evidence end-to-end conformance.
Why B is correct
System testing is defined as testing a complete, integrated system to verify that it satisfies its specified requirements, which is exactly what the question asks. Pilot testing checks whether a small user group can operate the system in a live setting and does not assert against specifications. Integration testing covers only the seams between components, and unit testing only the internal procedures of individual modules, so neither can confirm the application as a whole performs to specification.
Why the others are wrong
A. Pilot testing is a limited-scope user trial that validates operability and business fit, not specification conformance.
C. Integration testing only confirms that components interact correctly; components can pass it and still fail the specification.
D. Unit testing validates the internal control structure of a single module, far below the level of the whole application.
CISA exam tip
Conformance to the spec is a system-level claim: unit = module internals, integration = the seams, system = the whole thing to spec, pilot = can real users run it.
3While auditing a small organization's data classification processes and procedures, an IS auditor noticed that data is often classified at the incorrect level. What is the MOST effective way for the organization to improve this situation?
Conduct awareness presentations and seminars for information classification policies.
Use automatic document classification based on content.
Have IT security staff conduct targeted training for data owners.
Publish the data classification policy on the corporate web portal.
Answer: C
The short version
C — Targeted training for data owners fixes systemic misclassification. Owners decide handling, so their judgment drives labels. Focused coaching corrects root cause better than generic awareness or tools.
Key concepts in this question
Data classification: assigning sensitivity levels that trigger protection, retention and sharing rules.
Data owner role: business accountable party who understands value and risk, best placed to set correct levels.
Targeted training: role-specific guidance with examples from actual errors, closing observed knowledge gaps.
Generic awareness: broad seminars raise visibility but rarely change owner decisions on edge cases.
Automation limits: content tools help at scale yet still need correct policy interpretation by owners in small firms.
Why C is correct
C is correct because the auditor found frequent incorrect levels, indicating owners misunderstand criteria rather than lacking a portal link. IT security staff delivering targeted sessions can review mislabeled samples, clarify impact tiers and practice decisions relevant to that business. That direct accountability loop improves accuracy fastest and sustains compliance. Awareness decks, auto-labeling or publishing policy alone do not coach judgment where failures occur.
Why the others are wrong
A. offers general presentations, which inform broadly but miss the specific owner errors driving misclassification.
B. proposes automatic labeling, which needs tuning and oversight and does not teach owners to validate output.
D. publishes policy on a portal, improving access but not ensuring comprehension or consistent application.
CISA exam tip
Most effective for owner errors equals targeted training. Prefer accountability coaching over generic awareness or tooling alone.
4Which of the following should be the GREATEST concern for an IS auditor performing a post-implementation review for a major system upgrade?
Changes are promoted to production by the development group.
Developers have access to the testing environment.
Object code can be accessed by the development group.
Change approvals are not formally documented.
Answer: D
The short version
D — Change approvals are not formally documented. Without a documented approval trail there is no evidence any change was authorized, so the whole change control cycle is inoperative.
Key concepts in this question
Post-implementation review (PIR): An audit performed after go-live to confirm the delivered system matches what was authorized, funded and specified.
Change authorization: The first line of change management — every change must be requested, assessed for impact, approved by an authorized party, and the approval retained as audit evidence.
Audit evidence: Documentation is what makes an approval demonstrable after the fact; an oral or remembered approval cannot be relied upon or tested.
Why D is correct
Change management is ineffective if approvals are not documented, because the audit trail is the control. An undocumented approval gives an auditor no evidence that impact assessment, testing or authorization ever occurred, and leaves the organization exposed to unauthorized changes reaching production. The other observations are weaknesses that existing controls can still detect or mitigate: segregation of duties in promotion, developer access to a test environment, and developer access to object code are all conventional arrangements monitored by other controls.
Why the others are wrong
A. Developers promoting to production weakens segregation of duties, but it is a compensating control when approvals are documented and reviewed.
B. Developers having access to a testing environment is normal and expected; that is where development testing belongs.
C. Developers accessing object code is a normal part of the build process, not an exposure on its own.
CISA exam tip
If an approval cannot be evidenced, it did not happen: undocumented approval beats every other change weakness, because segregation of duties assumes authorization exists.
5Which of the following observations noted by an IS auditor reviewing internal IT standards is MOST important to address?
The standards have no reference to an industry-recognized framework.
The standards are not detailed in policies and procedures.
The standards are not readily available to organization-wide users.
The standards have not been revised within the last year.
Answer: A
The short version
A — The standards have no reference to an industry-recognized framework. Without a recognized framework the standards have no benchmark against which completeness can be judged.
Key concepts in this question
IT standards: The mandated rules for technology use (naming, architecture, security, data) that sit above procedures and below policy in the hierarchy.
Industry-recognized frameworks: COBIT for governance and control objectives, ISO/IEC 27001 and 27002 for security, NIST SP 800-53, COPC for process quality.
Benchmarking: A framework supplies the control objectives, so an auditor can test whether a standard is complete and can defend the organization against regulatory or third-party scrutiny.
Why A is correct
An IS auditor needs a baseline to evaluate anything. If internal standards reference no industry-recognized framework, the auditor cannot determine whether required topics, control objectives or regulatory drivers have been omitted, and cannot compare the organization to peers. The other observations are documentation, communication and maintenance-cycle gaps: they weaken the standards but do not remove the yardstick needed to know whether the standards are complete.
Why the others are wrong
B. Detail missing at the policy and procedure level is a granularity and enforcement problem, correctable by writing procedures.
C. Standards not widely distributed are a communication and awareness issue that can be remediated by publication and training.
D. An annual revision cadence is conventional practice, not a measure of the standard's substantive adequacy.
CISA exam tip
Framework = the yardstick. You cannot say what is missing from a standard set if nothing is measuring it.
6Which of the following is MOST important for an organization to consider when planning to outsource data storage to a third-party provider?
The cost of delivering the service
The country in which the provider operates
The classification levels of the stored data
The skill set and experience of the provider
Answer: B
The short version
B — The country in which the provider operates. The provider is governed by the law of its own jurisdiction, which overrides the organization's classification scheme.
Key concepts in this question
Data sovereignty: Where data is physically held determines which government, courts and regulators can compel access to it.
Outsourcing risk: Control is surrendered while accountability for privacy, retention and disclosure obligations stays with the organization.
Data classification: An internal label that drives handling controls; it binds the organization, not a foreign regulator or a provider acting under local law.
Why B is correct
Once data storage is outsourced, the organization's classification labels do not travel with the data. The provider must comply with the laws of the country where it operates, and those laws — privacy regimes, blocking statutes, disclosure orders — can conflict with the organization's obligations and classifications. Jurisdiction therefore imposes a risk the organization cannot mitigate through its own controls, which is why it outweighs cost, classification and provider skill set as the primary planning consideration.
Why the others are wrong
A. Cost is a legitimate commercial criterion, but an economic preference, not an uncontrollable exposure.
C. Classification levels determine the controls to apply, important but derivative once legal jurisdiction is set.
D. The provider's skill set affects service delivery quality, not the sovereign right to access the data.
CISA exam tip
Classification is your label; sovereignty is theirs. The host country's law binds the data no matter how you classified it.
7An IS auditor has been tasked with analyzing an organization's capital expenditures against its repair and maintenance costs. Which of the following is the BEST reason to use a data analytics tool for this purpose?
It reduces the sample size required to perform the audit.
It improves the reliability of the data.
It reduces the error rate.
It enables the auditor to work with 100% of the transactions.
Answer: D
The short version
D — It enables the auditor to work with 100% of the transactions. Full-population analysis removes sampling risk and tests every item.
Key concepts in this question
CAATs / data analytics: Tools that read complete data sets and apply rules across all records, rather than testing a sample.
Sampling risk: The risk that a sample does not represent the population, causing the auditor to miss misstatements or overstate exceptions.
Capital versus revenue: Capital expenditure is capitalized and depreciated; repairs and maintenance are expensed, and management has an incentive to misclassify between the two.
Why D is correct
The reason to reach for a data analytics tool is completeness of coverage. The auditor can extract and test every transaction classifying items as capital or revenue, so there is no sampling risk and no need to project findings. The tool does not repair the underlying data, does not make the population more reliable, and cannot reduce an error rate that the audit exists to measure; its value is the elimination of sampling.
Why the others are wrong
A. Sample size reduction understates the benefit, since the point is that the sample can be abandoned altogether.
B. Reliability depends on the source system's data quality, which the tool consumes unchanged.
C. The error rate is the subject under test; an analysis tool reports errors rather than preventing them.
CISA exam tip
Reach for analytics when you want 100% of the population, not a smaller sample and not cleaner data.
8Which of the following presents the GREATEST risk associated with end-user computing (EUC) applications over financial reporting?
Lack of portability for users
Calculation errors in spreadsheets
B. Inability to quickly modify and deploy a solution
Loss of time due to manual processes
Answer: B
The short version
B — Calculation errors in spreadsheets. Spreadsheets have no automated controls or audit trail, so a formula error goes straight into reported figures.
Key concepts in this question
End-user computing (EUC): Spreadsheets, databases and Access files built and maintained by business users outside the change, access and backup controls applied to formal applications.
Missing IT general controls: No segregation of duties over formulas, no change control, no version history, no review sign-off, no tested backup.
Financial reporting impact: EUC feeds directly into journal entries, reconciliations and management reporting, so an error is a misstatement rather than an operational annoyance.
Why B is correct
EUC risk to financial reporting is a data accuracy risk. Because the spreadsheet is an application without general controls, no automated test or review catches an incorrect or broken formula, and no version history shows who changed it. Calculation errors therefore reach the financial statements undetected, with direct financial and legal consequences. Portability and manual-process time losses are productivity concerns that do not change reported amounts.
Why the others are wrong
A. Lack of portability affects whether a solution can be moved or reused, not the accuracy of reported figures.
C. Lost time from manual processes is an efficiency issue with no direct effect on financial statement accuracy.
CISA exam tip
A spreadsheet is an application with no ITGC: no segregation of duties, no change control, no audit trail, so bad formulas reach the financials.
9An IS auditor should look for which of the following to ensure the risk associated with scope creep has been mitigated during software development?
Source code version control
Project change management controls
Existence of an architecture review board
Configuration management
Answer: B
The short version
B — Project change management controls. Scope creep is unapproved expansion of scope, and change control is the process that catches and authorizes it.
Key concepts in this question
Scope creep: Progressive, uncontrolled growth of project scope after the baseline is set, typically driven by unlogged requests.
Change management: A defined process — request raised, impact on schedule, cost and risk assessed, approved by the change control board, baseline and plan updated.
Configuration management vs. change management: Configuration management maintains the baseline and identity of items; change management authorizes movement of that baseline.
Why B is correct
Every item of scope creep begins as a change request that was never assessed, approved or reflected in the plan. Project change management controls force such requests through impact analysis and formal approval, so additions are either authorized and funded or rejected, keeping the project inside its baseline. The other options support development discipline but do not authorize scope changes, so none of them would surface an unlogged expansion.
Why the others are wrong
A. Source code version control tracks versions of code, not the project scope baseline.
C. An architecture review board advises on technology standards; it does not approve scope changes.
D. Configuration management maintains the baseline inventory, which does not by itself decide whether a change is accepted.
CISA exam tip
Scope creep is a change-control failure, never a version-control failure: every creep item is a change request nobody approved.
10Which of the following is MOST important to consider when defining disaster recovery strategies?
Mean time to restore (MTTR)
Maximum time between failures (MTBF)
Maximum tolerable downtime (MTD)
Mean time to acknowledge (MTTA)
Answer: C
The short version
C — Maximum tolerable downtime (MTD). MTD comes from business impact analysis and sets the outer limit that recovery objectives must meet.
Key concepts in this question
MTD: The maximum time a critical process can be unavailable before unacceptable business impact — the tolerance, not the plan.
RTO: The target time to restore a system; it must be shorter than MTD and is derived from it.
MTTR: The actual or expected time restoration takes; a performance measure to be minimized.
MTBF / MTTA: Mean time between failures is a component reliability metric; mean time to acknowledge measures detection and response speed.
Why C is correct
The strategy itself is dictated by how long the business can survive without the system, which is what MTD expresses and what business impact analysis establishes. Once MTD is known, RTO, RPO and the required redundancy and recovery approach follow from it. The other metrics describe how a system behaves or how fast people act, so they inform the plan but do not define the acceptable level of downtime.
Why the others are wrong
A. MTTR measures restoration effort after a failure; it is an outcome to minimize, not the tolerance that sets the strategy.
B. MTBF is a reliability statistic for components, unrelated to how much downtime the business can absorb.
D. MTTA covers detection and acknowledgement time, a response metric well below the tolerance threshold.
CISA exam tip
Business impact analysis gives you MTD; everything else (RTO, RPO, backups) must be built to fit inside it.