CISA Practice Test
Practise for the CISA with 715 exam-style questions, an explanation for every answer and stats that show what to study next.
What is the CISA exam?
The CISA exam is a 4 hour, 150-question multiple-choice test from ISACA, and you pass with a scaled score of 450 on a 200 to 800 scale. CISA stands for Certified Information Systems Auditor. Passing does not earn you CISA certification on its own: you also need five years of qualifying work, which you can finish after the exam.
Two domains make up over half the exam: Information Systems Operations and Business Resilience, and Protection of Information Assets, at 26 percent each. Plan your study around that. ISACA set August 1, 2024 as the start date for the exam on the current outline, and the ISACA outline page we read on November 7, 2026 showed no notice of a replacement.
Who can take the CISA exam?
You can register and take the exam before you have the experience. It counts when you apply for certification, not when you book the exam.
- You need five or more years of IS audit, control, assurance or security work, gained within the 10 years before you apply.
- Waivers can cover up to three years. Check the current application details on ISACA's site to see what qualifies for you.
- You must apply within five years of passing, so someone with two years of experience can take it now and still have time to reach five.
- Your supervisor or manager verifies your experience as part of the application.
CISA exam format and scoring
| Item | Official detail |
|---|---|
| Length | 150 multiple-choice questions |
| Time | 4 hours (240 minutes) |
| Question form | A stem and four options, with one best answer |
| Delivery | Computer-based, at PSI testing centers or remotely proctored |
| Passing score | Scaled score of 450, on a scale of 200 to 800 |
| Results | Preliminary pass status on screen, official score within ten working days |
Three scoring rules should change how you take the exam. Wrong answers carry no penalty, so never leave a question blank. Some items are unscored pretest items, and since ISACA's guide does not say how many and you cannot tell which they are, treat every question as if it counts. Domain results are for information only, because your score rests on total correct answers, whatever the domain.
The time works out at 1.6 minutes a question. Some questions hinge on a qualifier such as MOST likely or BEST, so several options may look defensible. Work out what the qualifier is asking before you compare the options.
What does the CISA exam cover?
The five domains and their weights come from ISACA's exam content outline. They differ in size and also in the kind of thinking they ask for.
- Domain 1, Information System Auditing Process (18 percent), is about how an auditor works: standards and ethics, risk-based planning, sampling, evidence and reporting. If you have never run an audit, expect to build this viewpoint from scratch.
- Domain 2, Governance and Management of IT (18 percent), moves up to structure, strategy, policy, enterprise risk, privacy and vendors. As you study it, keep asking who decides and who is accountable.
- Domain 3, Information Systems Acquisition, Development and Implementation (12 percent), follows a system from business case through design, testing, migration and post-implementation review.
- Domain 4, Information Systems Operations and Business Resilience (26 percent), covers daily running, such as incidents, change, patching, logs and service levels, along with business impact analysis, backups and recovery plans.
- Domain 5, Protection of Information Assets (26 percent), covers access, networks, encryption, cloud and mobile, then security testing, monitoring, incident response and evidence handling.
How should you split CISA study time?
Start with the weights, then adjust. ISACA's guide says each percentage is the share of the exam that reflects that domain. The hours below are arithmetic from those weights, not an ISACA recommendation, and we found no recommended study total from ISACA. The question counts are approximate too: ISACA does not publish a count per domain in anything we opened, and some items are unscored.
| Domain | Weight | Hours per 100 studied | Rough questions out of 150 |
|---|---|---|---|
| 1. Information System Auditing Process | 18% | 18 | 27 |
| 2. Governance and Management of IT | 18% | 18 | 27 |
| 3. Acquisition, Development and Implementation | 12% | 12 | 18 |
| 4. Operations and Business Resilience | 26% | 26 | 39 |
| 5. Protection of Information Assets | 26% | 26 | 39 |
Scale the hours to your own total. If you plan 150 hours, multiply each row by 1.5, which gives 39 hours each for Domains 4 and 5.
Then move time to suit your background. This part is our suggestion, not ISACA guidance. A security engineer might move some Domain 5 hours into Domains 1 and 2, while an accountant moving into IT audit might do the reverse. Only the total counts, so strength in one domain can offset weakness in another. Just avoid writing off Domains 4 and 5, which together are 52 percent of the exam.
If you are looking for a CISA study guide, ISACA's CISA page lists its own Review Manual, 28th Edition (2024). We did not assess third-party books for this page.
CISA costs, retakes and renewal
ISACA prices the exam in US dollars on the pages we read, so we do not convert it for India, the UK or anywhere else. A first-time pass costs US$625 for a member or US$810 for a nonmember. That covers the exam and application fees, but not membership dues, study materials or the yearly maintenance fee.
- Eligibility: you have six months from registering to take the exam. One six-month extension costs US$75.
- Booking: you can schedule from 48 hours after payment, and appointments open 90 days ahead.
- Retakes: you get four attempts in a rolling 12 months. The waits are 30 days, then 90, then 90, and each attempt costs the full fee.
- Rescore: if you do not pass, you can ask in writing within 30 days of results, for US$75.
- Renewal: at least 20 CPE hours a year and 120 over three years, plus the yearly maintenance fee.
Renewal rules change soon. For CPE reporting from January 1, 2027, at least 90 of the 120 CPE hours must align with the exam outline, and at most 30 can be other professional development. ISACA's CPE page has the details for your own cycle.
CISA practice test: what to do next
Read ISACA's outline, then take the 20 free questions on this page as a short mock test, timed at about 1.6 minutes each. Twenty questions is too few to judge a domain, so use them to learn the style. Read the explanation for every answer, including the ones you got right.
Each free question cites a public reference from NIST or GAO. Those are US government documents chosen for their subject content, so treat ISACA's outline as your guide to what the exam covers.
For more volume, VirtuePrep has 715 CISA practice questions, with an explanation for every answer. They are written for practice and are not real exam questions, and no practice score can promise a result. VirtuePrep is not affiliated with ISACA. You can also browse other exams.
20 free CISA practice questions
Answer each one yourself, then open the answer. No account needed. These are study questions, not real exam questions.
-
While on a six-month loan to the IT department, an internal IT auditor wrote the procedures for granting and removing access in a new ERP system. Back in audit, she is asked to lead an audit of those same controls because she knows the system well. What should she do FIRST?
- Accept the lead role and add a statement to the audit report saying she wrote the procedures
- Tell the chief audit executive about her earlier role before accepting the assignment
- Accept the lead role but limit her testing to whether staff have followed the procedures as written
- Ask the IT manager to confirm in writing that her procedures have been followed since she left
Show answer
Answer: B. She would be judging her own earlier work, which is a self-review threat. Threats to independence should be identified and evaluated first, which starts with telling the chief audit executive. Safeguards such as assigning the work to someone else follow from that evaluation. Disclosing the role only in the final report comes too late to protect the audit. The reference is a US government audit standard, and ISACA's Code of Professional Ethics also calls for objectivity.
Source: GAO, Government Auditing Standards, 2024 Revision, paragraphs 3.27 to 3.33 and 3.69
-
An audit team has hours for one more IT audit this year. Using a risk-based approach, which candidate should the auditor recommend?
- A warehouse scanning system, last audited five years ago against a three-year target, whose outages cause short shipping delays
- A data analytics platform holding no regulated data, which the CIO has asked audit to review because the project ran over budget
- A staff intranet with no sensitive content, which has drawn repeated complaints from senior managers about slow response times
- A supplier payment gateway that releases all vendor payments and moved to a new provider last quarter
Show answer
Answer: D. Risk is typically a function of the likelihood and impact of adverse events, so risk-based planning weighs both. The payment gateway has the highest impact, because it moves money, and the recent provider change can raise the likelihood of control failure. Time since the last audit, budget overruns and complaints may inform a plan, but none outweighs that combination here. The reference defines risk and is not an audit planning guide.
Source: NIST SP 800-30 Rev. 1, Guide for Conducting Risk Assessments
-
An auditor tests a random sample of 45 change tickets for documented approval. IT cannot find two of the selected tickets, no other record can show that those changes were approved, and the result for those two items could change the audit conclusion. How should the auditor treat these two items?
- Replace them with two more randomly selected tickets and test those instead
- Remove them from the sample and evaluate the results on the remaining 43 tickets
- Count them as deviations in the evaluation and consider why they are missing
- Accept a signed statement from the change manager that the two changes were approved
Show answer
Answer: C. When unexamined items could alter the conclusion, the auditor tries alternative procedures. If those cannot be completed, the items count as deviations in the evaluation, and the auditor also considers what the missing records may imply. Swapping or dropping the items would hide a possible control failure. A statement written afterward is weak evidence that approval happened. The reference is a financial audit manual, but the same logic applies when you sample a control in an IT audit.
Source: GAO and CIGIE, Financial Audit Manual, Volume 1 (June 2026), section 450.20
-
An auditor needs to determine whether employees who left the company still have access to the finance application. Which evidence is MOST reliable?
- A user list the auditor watched being extracted from production, then matched to HR termination records
- A spreadsheet prepared by the application administrator showing the date each former employee's account was removed
- A signed statement from the IT director that access is removed within one business day of an employee leaving
- A copy of last year's external audit report that describes the offboarding process as operating effectively
Show answer
Answer: A. Evidence the auditor obtains through direct observation, inspection and computation is generally more reliable than evidence obtained indirectly. The other options are prepared by others, rest on what someone says, or describe an earlier period. The reference is the performance audit chapter of a US government audit standard.
Source: GAO, Government Auditing Standards, 2024 Revision, paragraph 8.104b
-
A board-approved risk statement says no internet-facing system may carry a known critical flaw for more than 14 days, and only the CIO may approve exceptions. An auditor finds that an IT manager signed a one-year exception for the customer portal, and no one above him was told. What should the auditor view as the MOST significant concern?
- An IT manager accepted a risk above the stated tolerance without the authority to do so
- The exception form does not record the severity score that was assigned to the flaw on the portal
- A one-year exception period is longer than is normally needed to plan a repair
- The portal remains on infrastructure that makes this type of flaw slow to repair
Show answer
Answer: A. Risk tolerance statements are established and communicated so that decisions stay within them, and leadership remains accountable for cybersecurity risk. A manager quietly going beyond that tolerance is a governance failure: the people who own the risk do not know they are carrying it. The other points may need attention, but they do not address who made the decision.
Source: NIST Cybersecurity Framework 2.0, categories GV.RM and GV.RR
-
A retailer's policy says customer records are deleted 24 months after the last purchase. The CRM runs a deletion job, but an auditor finds customer data also feeds an analytics platform and a marketing tool, and the team cannot say what other systems receive it. What should the auditor recommend FIRST?
- Extend the CRM deletion job to the analytics platform and the marketing tool
- Shorten the retention period to 12 months to reduce the data held across systems
- Create and maintain an inventory and map of the systems, owners and data involved
- Encrypt all of the customer data held in the analytics platform and the marketing tool
Show answer
Answer: C. The organization cannot show that data is destroyed according to policy until it knows where the data goes. An inventory and map of systems, owners, data elements and purposes gives that view. Extending the job to the two known systems could still miss others, and encryption or a shorter period does not fix the blind spot.
Source: NIST Privacy Framework 1.0, category ID.IM-P and subcategory CT.DM-P5
-
A payroll provider has moved customer support to a subcontractor that can view employee records. The company's contract sets security requirements for the provider but does not mention subcontractors, and the company last checked the provider's security when it signed three years ago. What is the BEST recommendation?
- Terminate the contract and bring payroll processing back in house as soon as possible
- Amend the contract to extend requirements to subcontractors and recheck compliance periodically
- Obtain a one-time written confirmation from the provider that the subcontractor is secure
- Ask the subcontractor each year for copies of the security certificates listed on its public website
Show answer
Answer: B. Contract terms should carry requirements down to subcontractors where they apply and provide for periodic checks that suppliers still meet them. That closes both gaps. A one-time letter gives no ongoing check, a certificate the subcontractor chooses to publish creates no contractual duty, and terminating the contract goes further than the facts support at this stage.
Source: NIST SP 800-161 Rev. 1 (updated 2024), section 3.1.2
-
An IT security strategy sets the objective of shortening the time taken to detect and contain security incidents. The monthly scorecard to executives reports only tickets closed and training completed, and both are consistently green. What should the auditor recommend?
- Raise the targets for tickets closed and training completed so the scorecard is more demanding
- Replace the monthly scorecard with a yearly maturity rating from an outside assessor
- Transfer ownership of the scorecard to internal audit so that its results are independent
- Add measures tied to the objective and have leadership use the results to adjust the strategy
Show answer
Answer: D. Reporting helps oversight when it shows progress against the strategy, because leaders can then use the result to adjust direction. Green activity counts say nothing about the stated objective. Moving the scorecard to audit would impair audit's objectivity, and a yearly rating is too infrequent to steer the work.
-
A team building a customer portal is finalizing the design of its payment and login features. Its only planned security activity is a penetration test one week before go-live. Which recommendation would MOST reduce the cost of security defects?
- Carry out threat modeling now and feed the results into the design and requirements
- Add a second penetration test by a different firm during the same week before go-live
- Require the CISO to sign the penetration test report as a condition of go-live
- Schedule a code review of the payment and login features after the first release
Show answer
Answer: A. Assessing security risk during design lets the team address weaknesses before they are built. NIST's secure development framework says that addressing security earlier takes less effort and cost, and that handling risks during design also helps development efficiency. The other options all find or sign off on problems late, after the design is set.
Source: NIST SP 800-218, Secure Software Development Framework 1.1, section 1 and practice PW.1
-
Before cutover to a new claims system, the migration team loads a full copy of production claims into a test environment so users can check the conversion. The test environment has wider access than production, and no one approved this use of the data. What is the BEST recommendation?
- Delete the copied data from the test environment within 30 days after go-live
- Have every tester sign a confidentiality agreement before they receive access to the test environment
- Move all conversion checks to production and carry them out after the cutover
- Approve the data use, protect the test environment to suit the live data, and prefer masked data
Show answer
Answer: D. Live data in preproduction environments needs approval and control, and the environment should be protected at the same level as the data it holds. Using test or dummy data where possible reduces the risk further. Agreements and later deletion leave the too-wide access in place, and testing only after cutover defeats the point of a rehearsal.
-
A furnace controller at a manufacturer runs an operating system its vendor no longer patches, and it cannot be replaced for three years. In 2023 IT recorded the risk as accepted and took no further action. The controller still sits on the general corporate network. What should the auditor recommend?
- Install the latest available patches for a similar operating system version on the controller
- Renew the risk acceptance each year with a signature from the plant manager
- Isolate the controller from other systems and reassess the mitigation periodically
- Bring forward the replacement by moving funds from other planned plant projects
Show answer
Answer: C. Assets that cannot be patched still need long-term mitigations, such as isolation from other assets. Those mitigations should be reviewed periodically through risk assessment and cost-benefit analysis. Signing the acceptance again leaves the exposure unchanged, and patches meant for another version could disrupt a process controller. The facts given rule out early replacement.
-
Investigators compare firewall, application and database logs after a data theft. A database export appears about 90 seconds before the login that caused it. They find each server sets log timestamps from its own clock. Which control would BEST have prevented this problem?
- Keeping logs on each server for a longer retention period
- Synchronizing every logging host to a common time source
- Raising each server's logging to its most detailed setting
- Restricting who is allowed to read the logs on each server
Show answer
Answer: B. Each host stamps entries using its internal clock, so inaccurate clocks can put events in the wrong order when logs from several hosts are compared. Syncing clocks to a common source fixes that directly. More retention, more detail or tighter access would not correct the timestamps.
Source: NIST SP 800-92, Guide to Computer Security Log Management, sections 2.3.1 and 5.5
-
During an active attack on a Saturday night, an engineer added a firewall rule through the emergency change procedure, before the change board could meet. An auditor is testing this change. Which evidence would BEST show that the emergency process worked as a control?
- Records of expedited authorization, then after-the-fact analysis, testing and approval of the rule
- A report showing that the attack stopped soon after the engineer had added the new firewall rule
- An oral confirmation from the engineer's supervisor that the rule was clearly necessary at the time
- A firewall configuration export showing that the new rule remains in place on the device today
Show answer
Answer: A. Unscheduled changes can follow expedited reviews and approvals, but they should still receive analysis, testing and approval after the fact, including a look at any effect on security controls. That record shows the process operated. The other options show the rule existed or seemed to help, not that the change was controlled.
-
A business impact analysis sets a maximum tolerable downtime of 8 hours for order processing. The recovery plan restores the order system in 6 hours, and staff then need 3 hours to reprocess the orders affected. Hourly backups meet the recovery point objective. What should the auditor conclude?
- The plan is adequate because the 6 hour recovery time is shorter than the 8 hour limit
- The plan is adequate because hourly backups keep any data loss within the agreed objective
- The plan falls short because restoration plus reprocessing takes 9 hours against an 8 hour limit
- The plan falls short because the recovery point objective should be added to the recovery time target
Show answer
Answer: C. The recovery time objective plus the time to reprocess work has to fit within the maximum tolerable downtime, and 6 plus 3 gives 9 hours against a limit of 8. The recovery point objective concerns data loss and is not part of the downtime calculation.
-
Backup logs for a core system show 100 percent successful jobs for the past year, and backup media go to an off-site vendor each week. Which procedure would give an auditor the BEST assurance that the system can be recovered?
- Comparing each job in the backup logs with the approved backup schedule for the year
- Reviewing the off-site vendor's receipts to confirm that every weekly delivery was made
- Obtaining the backup administrator's written confirmation that all of the jobs are monitored daily
- Observing a restoration from off-site media, with the system owner checking the restored data
Show answer
Answer: D. A job can report success and still leave data that cannot be retrieved, so backups should be tested to show files can be read back without errors or loss, including at an alternate site where one applies. Logs, receipts and confirmations show the routine ran, not that recovery works.
-
Attackers set up a look-alike admin portal and collected passwords and one-time codes from several administrators, then relayed them to the real portal at once. Which change would MOST directly stop this attack from working?
- Sending the one-time codes by text message rather than delivering them through an authenticator app
- Shortening the validity of the one-time codes from 60 seconds to 30 seconds
- Issuing administrators cryptographic security keys that require no typed code
- Giving administrators printed look-up codes to use in place of the one-time codes
Show answer
Answer: C. Codes that a person types in are not bound to the session, so an impostor site can relay them to the real one. Phishing resistance requires cryptographic authentication that binds the login to the session or the real site's name, so it does not rely on the user spotting the fake site. Changing how a code is delivered, shortening its life, or printing codes still leaves a typed secret to relay.
Source: NIST SP 800-63B-4, section 3.2.5
-
A document system uses one RSA key pair both to sign approvals and to decrypt the keys that protect archived files. The team plans to keep the private key indefinitely so that old archives stay readable. What is the MOST significant weakness in this design?
- Keeping the key for archive access conflicts with destroying a signature key at end of cryptoperiod
- RSA key pairs can be used to sign data only, so the archive decryption will eventually fail
- The public key would have to be reissued whenever a new archived file was encrypted
- Decrypting archived files with the private key would invalidate every approval that was signed earlier
Show answer
Answer: A. A single key should generally serve one purpose, and this is the kind of conflict NIST describes. A key used for key transport may need to be kept long after its public key expires, but a private signature key should be destroyed at the end of its cryptoperiod. The system should use separate key pairs.
-
A finance application runs on virtual machines rented under an infrastructure as a service agreement. An auditor finds the guest operating systems are missing security patches, and IT says the provider handles patching. Nothing in the contract adds operating system management to the service. What should the auditor conclude?
- The provider is at fault because it owns and runs the servers beneath the virtual machines
- The company is responsible because the consumer controls the operating systems under this model
- Responsibility is shared equally because neither party has full control over the virtual machines
- No finding is needed because the risk of unpatched systems transferred under the agreement
Show answer
Answer: B. Under infrastructure as a service, the provider manages the underlying cloud infrastructure, while the consumer controls operating systems, storage and deployed applications. With no contract term moving that duty, IT's assumption is wrong and the patch gap is the company's to fix.
Source: NIST SP 800-145, The NIST Definition of Cloud Computing
-
A network vulnerability scan reports 400 critical findings, and administrators suspect many are false positives. The security team plans to run a second vendor's scanner and treat every finding reported by both as confirmed. What should the auditor recommend for validating the findings?
- Keep the plan, because agreement between two independent tools removes nearly all the false positives
- Run the original scanner again outside business hours and compare the two sets of results
- Accept the risk ratings assigned by the scanner and send all 400 findings for remediation
- Have a qualified assessor manually examine affected systems, starting with the highest risk
Show answer
Answer: D. Scanners can have high false positive rates. NIST accepts a second tool as one way to validate findings, but comparison tools can repeat the same false positives, so agreement alone does not confirm a finding. Manual examination is typically more accurate but takes more time, so it makes sense to start with the highest-risk systems. Assessors should interpret scanner results and not simply accept assigned risk levels.
-
A server suspected of compromise is still running. Legal says the evidence may be needed for disciplinary proceedings. An administrator proposes rebooting it to remove the malware and then taking a disk image. What should be done FIRST?
- Run a full antivirus scan on the server so the malware can be identified before anything else changes
- Collect volatile data from the running server, under chain of custody, before any reboot
- Disconnect the power cable immediately so that the disk contents cannot be altered any further
- Reboot as proposed, because the disk image will preserve everything needed as evidence
Show answer
Answer: B. Volatile data exists only on a live system and is lost on shutdown or reboot. NIST's guide says that in most cases you collect network connections and login sessions first. Because proceedings are possible, chain of custody should be followed. Scanning alters the system, and removing power or rebooting destroys volatile data.
What you get with VirtuePrep CISA prep
- 715 practice questions built around the topics the CISA tests.
- An explanation for every answer, so each mistake teaches you something.
- Stats and history that show your weakest topics first.
- Web, iPhone and Android, so you can study wherever you have ten minutes.
CISA practice test questions people ask
What score do you need to pass CISA?
What is the CISA passing percentage?
What is the CISA pass rate?
How much does the CISA exam cost?
What is the CISA exam format?
How hard is the CISA exam?
Can I take CISA practice questions for free?
Sources
All pages were opened on November 7, 2026.
- ISACA: CISA certification
- ISACA: CISA exam content outline
- ISACA: CISA exam information
- ISACA: 2026 exam candidate guide (PDF)
- ISACA: Get CISA certified
- ISACA: Maintain CISA certification
- ISACA: CPE changes for 2027
- ISACA: 2024 press release on the updated CISA exam
- ISACA: Free CISA practice quiz
VirtuePrep is not affiliated with or endorsed by the organisation that runs the CISA. Exam names are trademarks of their owners. Practice questions are written for study and are not real exam questions. Page updated 2026-11-07.
More Security exams
CISM
The CISM exam isn't just another certification test, it measures how you think and make decisions as a security leader.
SecurityCISSP
The CISSP exam is one of those tests that looks fine on paper until you actually start studying...
SecurityCompTIA Security+
The Security+ exam covers a lot more than people expect the first time they look at it.
SecurityCRISC
The CRISC exam can feel confusing when you first start because it mixes risk, controls, business processes, and IT...
Start your CISA prep today
Answer your first question in under a minute. Free to try.