ISC · Security and Confidentiality and Privacy · 6 practice questions
Data Minimization for Vendor Testing and Analytics
Share only the data needed for the task, and mask unnecessary customer identifiers before granting access. Below, follow a vendor dashboard test from the initial data request to a limited, protected test dataset.
Try one first
Hint
Focus on the control that reduces exposure of personal data before it leaves the company, not just the controls applied after sharing.
Answer D. This answer applies data minimization and de-identification: share only the attributes required for the test and remove or mask direct identifiers to reduce unnecessary exposure of personal information. Management approval and confidentiality agreements are important but do not justify sending unneeded PII; minimizing the dataset before it leaves the company best supports confidentiality and privacy.
Why not A: Tempting because encryption and contracts are valid controls, but they do not remove the privacy risk of sharing unnecessary personal data. Sending the full production file exposes unneeded identifiers and increases breach risk.
Why not B: While consent can be important, requiring fresh consent from all customers is an extreme next step given the facts. The immediate better control is to minimize and de-identify the data shared; obtaining fresh consent is not the best or most practical next action presented here.
Why not C: Read-only production access may seem efficient but still exposes unnecessary data and broadens system access. It increases operational risk by granting the vendor access to live systems rather than limiting the dataset to what the vendor actually needs.
Step by step
- Define the vendor's task
A company hires a vendor to test a dashboard using customer ID, region, and purchase totals. Production also contains names, Social Security numbers, full birth dates, and street addresses; the vendor has signed confidentiality terms, and encrypted transfer is available.
- Separate relationships from identities
The dashboard needs to link purchases to customer records, not identify actual customers. Preserve those links with consistent masked customer IDs rather than exposing the original IDs.
- Remove every unnecessary field
Build an extract containing only the masked customer ID, region, and purchase totals needed for testing. Exclude names, Social Security numbers, full birth dates, and street addresses.
- Reject full-file shortcuts
Do not send the full production file just because encryption and confidentiality terms are available. Read-only production access also fails the test because the vendor could still view unnecessary personal data.
- Protect the reduced dataset
Use an approved secure transfer method and limit access to authorized vendor testers in the test environment. Masking reduces exposure but does not eliminate the need to protect the remaining data.
- Control any live-data exception
If testing later requires identified data, require a documented need and privacy/security approval before access. Limit access to necessary fields, records, and users, make it time-limited, and log its use.
Key points
- Remember: protecting a full customer file does not justify sharing fields the vendor does not need.
- Preserving links between test records does not require revealing customer identities.
- Read-only access still exposes every personal field the vendor can view.
How the exam traps you
- Send the full file because encryption and confidentiality terms are in place. Minimize first, then protect the reduced dataset during transfer and use.
- Remove names but leave unnecessary Social Security numbers, birth dates, and addresses. Review every field against the testing purpose, not just the name field.
- Give read-only access to production instead of preparing an extract. Provide a limited test dataset; read-only access does not prevent unnecessary disclosure.
Question 2
Hint
Focus on which security objective is most directly affected when users without a business need can no longer see full sensitive data.
Answer B. Masking full card numbers for users who don't need them is a least-privilege and data-minimization control that reduces the risk of unauthorized disclosure of sensitive information. Because the stem specifies that processing and uptime are unchanged, the primary security objective affected is confidentiality/privacy rather than integrity or availability.
Why not A: This is tempting if a student conflates visibility with correctness, but data integrity concerns the accuracy and completeness of data and validation controls, not whether a user can see the full value. Masking affects exposure, not the underlying accuracy of stored or processed values.
Why not C: This overgeneralizes the benefit of masking in one interface. Full card data still exists for billing and may be stored in databases, logs, or backups, so those copies must remain protected; masking narrows exposure in the UI but does not remove other protection obligations.
Why not D: Availability refers to access to systems and data when needed, not how widely sensitive fields are displayed. Restricting who can view full card numbers primarily affects confidentiality, not system availability.
Question 3
Hint
Focus on the difference between protecting data that is needed and reducing collection of data that is not needed.
Answer C. The most direct way to reduce privacy risk is to minimize the collection and retention of personal data that is not required for the app's purpose. While encryption and access controls protect stored data, they do not eliminate the risk caused by holding unnecessary exact geolocation data. Stopping collection and deleting unneeded records reduces exposure and compliance risk.
Why not A: Confidentiality agreements help deter and address misuse by personnel, but they do not reduce the underlying privacy risk created by collecting and retaining unnecessary personal data.
Why not B: Stronger authentication improves security, but it doesn't address the core privacy issue here: the company is collecting exact geolocation data that it doesn't need. Better security doesn't remove unnecessary data.
Why not D: A backup site enhances availability and resilience but creates additional copies of data, which can increase the scope of what must be protected and does not reduce privacy risk from over-collection.
Question 4
Hint
Focus on the control that reduces privacy risk before the data is stored, not just the control that protects data after it has already been collected.
Answer C. Data minimization and defined retention limits reduce privacy risk at the source by avoiding unnecessary collection and limiting how long personal data is held. Given the stated purpose (promotions and reward balances) and no legal or fraud-control need for DOB or SSN, the company should collect only what is required and delete or anonymize data when no longer needed. This approach is a stronger privacy control than security or administrative safeguards applied after excessive data has been collected.
Why not A: Frequent password changes address account security, not the core privacy issue of collecting or retaining more personal data than necessary. Even with password rotation, unnecessary data still increases privacy risk.
Why not B: Encryption protects confidentiality from external disclosure but does not eliminate the privacy harm of collecting data that the business does not need. Minimizing collection and retention reduces exposure more effectively than securing unnecessary data after collection.
Why not D: Confidentiality agreements are a useful administrative control but do not prevent the business from collecting or retaining unnecessary personal data. They reduce insider misuse risk but do not address the root cause of excess data collection.
Question 5
Hint
Distinguish confidentiality controls (encryption, access limits) from privacy governance (what data are collected and whether those data are needed for the stated purposes).
Answer B. Encryption and need-to-know access are confidentiality (security) controls that reduce unauthorized access. Privacy also concerns whether personal data are collected and used appropriately for stated purposes. Because the retailer requires dates of birth but does not use them for processing orders or administering rewards, unnecessary collection creates a privacy issue even though confidentiality improved.
Why not A: Tempting because encryption and access limits do reduce misuse and are helpful to privacy outcomes; however, they are confidentiality controls. Privacy additionally covers data minimization and purpose limitation, and collecting unnecessary dates of birth remains a privacy concern despite stronger security.
Why not C: Tempting because notice (disclosure) is a privacy practice and the form states purposes; but disclosure alone does not make collection appropriate. Collecting personal data not needed for the stated purposes still violates privacy principles even if the company provided notice.
Why not D: Tempting because external disclosure is a major privacy risk and keeping data internal reduces that risk; however, privacy also includes whether the company should collect particular personal data. Retaining unnecessary personal information is a privacy issue even if it is not shared externally.
Question 6
Hint
Which choice gives the team the specific data elements they said they need while exposing the least personal information?
Answer A. The team needs diagnosis codes and dates of service but does not need identifiable information. Company policy requires the minimum necessary data and favors lower-risk alternatives for testing, so producing a masked or de-identified dataset that retains only required fields and restricting access satisfies the business need while minimizing privacy and re-identification risk.
Why not B: This is tempting because a documented exception plus encryption and access controls reduces some operational risk and preserves full realism for testing. It is still wrong here because it exposes unnecessary identifiers contrary to the minimum-necessary policy when a lower-risk alternative (masked/de-identified data) is practical.
Why not C: Asking for consent is strongly protective of privacy and may be appropriate in some contexts, which makes this option appealing. It is incorrect here because it is an over-restrictive response that is unnecessary when a de-identified alternative can meet the testing need and would unduly delay a legitimate internal project.
Why not D: This seems attractive because pseudonymization and removing SSNs preserve realism while adding controls. It fails because remaining quasi-identifiers (addresses, dates of birth, diagnosis codes) can still enable re-identification, and the policy calls for minimizing personal data exposure when a lower-risk masked dataset will suffice.
Common questions
Does encryption replace data minimization for a vendor?
No. Encryption protects data during transfer and storage, but it does not justify giving the dashboard vendor unnecessary customer identifiers.
Should customer IDs be masked in test data?
If the dashboard needs record relationships rather than actual identities, use consistent masked IDs to preserve those relationships. Do not expose actual customer IDs merely for convenience.
Can live customer data ever be used for testing?
Yes, if a specific need for identified data is documented and approved. Limit the fields, records, users, and access period, and log access.
Watch it solved
A full CPA ISC task-based simulation on Confidentiality and privacy, worked step by step.
Data Classification & Privacy Compliance | CPA ISC Simulation (TBS) on YouTube
Practice ISC like the real exam
The free ChatCPA simulator: real exam layout, timed testlets, starting with a question on this topic. No account needed to start.
Open the free simulator →More on Confidentiality and privacy
Questions from the ChatCPA bank of 17,658 CPA exam questions, each with a written reason for every wrong answer. ChatCPA is built by Nicholas Miller, CPA (Oregon #14907). How these pages are made. Spot an error? Tell us.