Below are 10 original CIPT-style practice questions, each with a hidden answer and explanation. They cover minimisation, de-identification (hashing, k-anonymity, l-diversity), differential privacy, encryption and key management, privacy by default, dark patterns, deletion and privacy threat modelling. The live exam is 90 questions in 2.5 hours, needs a scaled 300 of 500, and follows BoK v4.0.0. Score yourself with the table at the end.
The CIPT is the IAPP's most technical exam. Candidates from law and policy usually lose marks not on privacy principles but on how privacy is engineered: which de-identification technique fixes which weakness, and what a parameter actually does. These ten questions test exactly that. If your exam is booked, use them to find the technical gaps while there's still time to close them.
How to use this set
- Answer before you open the explanation. Commit to a letter and a reason. Right answer with the wrong reason counts as a miss.
- Name the technique and what it protects against. If you can't, you guessed well rather than knew it.
- These are original questions written in the exam's style, not real exam items.
Practice questions
A streaming service only needs to confirm that new users are 18 or over. Product wants to collect full date of birth "in case marketing wants birthdays later." What is the best privacy engineering choice?
- ACollect full date of birth and encrypt it at rest
- BCollect date of birth and delete it after five years
- CCollect date of birth but restrict access to the marketing team
- DVerify age at sign-up and store only an over-18 flag
Show answer
Correct: D
Data minimisation means collecting only what the stated purpose needs. Here that's proof of age, not the birthday. Encryption (A) and access control (C) are security measures. They protect data you shouldn't hold in the first place. A retention limit (B) shortens the problem without fixing it. "In case marketing wants it later" is not a specified purpose.
Domain II: collection and minimisationA team shares unsalted SHA-256 hashes of customer email addresses with an advertising partner and calls the data "anonymised." What is the best assessment?
- ACorrect, because hashing is irreversible
- BCorrect, as long as the partner signs an NDA
- CIncorrect. SHA-256 is broken and should be replaced with MD5
- DIncorrect. The partner can hash its own email list and match records, so the data is pseudonymised and still linkable
Show answer
Correct: D
A hash of a known identifier is a stable pseudonym. Anyone holding the same emails can hash them and join the datasets, so the data stays linkable to people. That's pseudonymisation, not anonymisation. An NDA (B) is a contract, not a technical property. MD5 (C) is weaker, not stronger.
Domain II: de-identificationA health dataset is 5-anonymous on ZIP code, age band and sex. In one equivalence class, every record has the same diagnosis. Which technique addresses this weakness?
- AIncreasing k to 10
- BEncrypting the diagnosis column at rest
- CRemoving patient names
- Dl-diversity
Show answer
Correct: D
This is a homogeneity attack. Even without knowing which record is theirs, anyone who can place a person in that group learns the diagnosis. l-diversity requires several distinct sensitive values in each group (t-closeness goes further). A larger k (A) doesn't guarantee diversity, encryption at rest (B) doesn't change what authorised users can infer, and names (C) are direct identifiers that should already be gone.
Domain II: de-identification modelsAn analytics team lowers the differential privacy parameter epsilon (ε) from 1.0 to 0.1. What is the effect?
- ALess noise, weaker privacy, more accurate results
- BMore noise, stronger privacy, less accurate results
- CNo change to privacy; only performance improves
- DThe data becomes encrypted
Show answer
Correct: B
Epsilon is the privacy loss budget. A smaller ε means more calibrated noise, a stronger guarantee that one person's data barely changes the output, and lower accuracy. Choosing ε is a privacy-utility trade-off, which is exactly the kind of judgement the CIPT tests. Differential privacy is not encryption (D).
Domain V: privacy-enhancing technologiesA customer database is encrypted at rest, but the encryption keys are stored in a configuration file on the same server. What is the main weakness?
- ANone. Encryption at rest is sufficient
- BAn attacker who compromises the server gets both the data and the keys; keys should be held separately, for example in a KMS or HSM, with access control and rotation
- CAES is too weak for personal data
- DThe keys should be emailed to administrators as a backup
Show answer
Correct: B
Encryption is only as strong as its key management. Keys stored next to the data mean one compromise unlocks everything. Separate key storage (a KMS or HSM), least-privilege access and rotation fix that. AES (C) is not the problem, and emailing keys (D) makes it worse.
Domain III: risk and controlsWhich approach is most appropriate for storing user passwords?
- AUnsalted SHA-256 hashes
- BReversible AES encryption so support staff can read passwords
- CMD5 with a shared salt
- DA salted, deliberately slow adaptive hash such as Argon2 or bcrypt
Show answer
Correct: D
Passwords should be stored with a salted, slow, adaptive hash built for the job (Argon2, scrypt, bcrypt). Unique salts defeat precomputed tables, and slowness makes brute force expensive. Fast hashes (A, C) are easy to crack at scale. Reversible encryption (B) means anyone with the key can read every password.
Domain III: controlsA new social app is deciding how location sharing should work. Which design best reflects privacy by default?
- ALocation sharing on, with an opt-out in settings
- BLocation sharing on, but stored encrypted
- CAsk for permission at every login
- DLocation sharing off until the user turns it on, with a clear explanation of who will see it
Show answer
Correct: D
Privacy by default means the most protective setting applies without any action from the user. The user then chooses to share, knowing the consequences. An opt-out default (A) puts the burden on the user, encryption (B) doesn't change who sees the location, and constant prompts (C) cause consent fatigue, which is its own usability failure.
Domain IV: privacy by designA cookie banner shows a large "Accept all" button. Rejecting non-essential cookies takes three screens through "Manage settings." What is the best fix?
- AMake the "Accept all" button smaller
- BOffer a "Reject all" option on the first layer with the same prominence as "Accept all"
- CAdd more categories to the settings screens
- DPre-tick all the optional categories to save the user time
Show answer
Correct: B
This is an asymmetric choice / obstruction dark pattern: accepting is easy and refusing is hard. The fix is equal ease and prominence for both choices on the first layer. Shrinking a button (A) doesn't remove the obstruction, and pre-ticked boxes (D) are another dark pattern that doesn't produce valid consent.
Domain IV: UX and dark patternsA user's deletion request is carried out in production, but their data remains in nightly backups retained for 90 days. What is the most appropriate practice?
- ARestore and edit every backup immediately
- BIgnore backups, because they're out of scope for deletion
- CDocument the backup retention period, make sure deleted records are suppressed if a backup is ever restored, and let backups age out
- DKeep backups forever in case the user returns
Show answer
Correct: C
Editing backups (A) is usually impractical and risky. The accepted approach is a documented, limited backup retention plus controls that stop deleted data reappearing, such as a deletion log or suppression list re-applied on any restore. Backups are not out of scope (B), and indefinite retention (D) breaks retention limits.
Domain II: retention and destructionAn engineering team already uses STRIDE for security threat modelling. Which framework is designed specifically to model privacy threats?
- ALINDDUN
- BDREAD
- COWASP Top 10
- DCVSS
Show answer
Correct: A
LINDDUN names privacy threat categories: Linkability, Identifiability, Non-repudiation, Detectability, Disclosure of information, Unawareness and Non-compliance. It's used the same way STRIDE is for security. DREAD and CVSS (B, D) score severity, and the OWASP Top 10 (C) lists common web application security risks.
Domain III: privacy risk modelsHow to read your score
| Score | What it means |
|---|---|
| 9–10 with the right reasons | Your CIPT knowledge is in good shape; move to full timed sets |
| 6–8 | Typical a few weeks out; your misses show you what to review |
| 5 or fewer | Go back to the material before more questions. Focus on de-identification and PETs first; they're the biggest gap for non-technical candidates. |
What your misses usually mean
- Missed Q2 or Q3? De-identification is your gap. Learn the difference between pseudonymised and anonymised data, and what each model (k-anonymity, l-diversity, t-closeness) defends against.
- Missed Q5 or Q6? You're treating "encrypted" or "hashed" as automatically safe. The exam tests how it's done.
- Missed Q7 or Q8? Review privacy by design and UX. These are often the easiest marks for technical candidates to drop.
Next: the CIPT exam guide, how many hours to study and CIPT vs CIPM.
Frequently asked questions
Are these real CIPT exam questions?
No. They're original questions written in the style of the exam. The IAPP doesn't release live exam items.
How many questions are on the CIPT exam?
90 multiple-choice questions in 2 hours 30 minutes, plus an optional 15-minute break. You need a scaled 300 out of 500 to pass. The current Body of Knowledge is v4.0.0, effective 1 September 2025.
Is the CIPT very technical?
It's the most technical of the core IAPP exams. You don't write code, but you need to understand how de-identification, encryption, privacy-enhancing technologies and privacy by design work well enough to choose the right one for a scenario.
I'm not an engineer. Can I pass the CIPT?
Yes, but budget more time. Non-technical candidates usually need the upper end of the 30–60 hour range, and should focus practice on de-identification, PETs and security controls.
Sources
- IAPP: CIPT certification
- LINDDUN privacy threat modelling framework
- NIST SP 800-188: De-identifying government datasets
- OWASP Password Storage Cheat Sheet
Last reviewed: November 2026.