These 12 CBDA source data questions cover Domain 2 of the IIBA blueprint: planning data collection, choosing data sets, collecting and validating data, and the techniques behind that work. Each is an original scenario with one best answer, except one "choose two" item, and the key cites the Guide section for every answer. If your research question is not framed yet, start with moving from a business problem to a research question.
How this cbda source data questions set is built
Source Data carries 15% of the exam. The Guide to Business Data Analytics treats sourcing as top-down: understand the problem context first, then decide what data to use (2.2). The business analysis professional adds the meaning behind variables that a pure data view can miss.
The four tasks in brief:
- 2.2.1 Plan Data Collection. What data is needed, whether it is available, whether history is required, when and how to collect it, and how to validate it. Non-functional needs such as privacy, retention, volume and frequency shape the plan.
- 2.2.2 Determine the Data Sets. Data types, dimensions, sample size, relationships, whole versus partial sets, and data gaps. A five Vs assessment and a cost-benefit view decide what to keep; profiling and sampling make large sources manageable.
- 2.2.3 Collect Data. Test the approach on a few observations first, trace data lineage, and reconcile sources that encode the same thing differently.
- 2.2.4 Validate Data. Business validation (stakeholders approve sources and set acceptance criteria) plus technical validation of accuracy, completeness, consistency, uniqueness and timeliness.
The set also uses five techniques listed in 2.2.5: Data Dictionary (3.4), Data Flow Diagrams (3.5), ETL (3.10), Interface Analysis (3.13) and Survey and Questionnaire (3.17).
Each question has a short scenario and a stem asking for the BEST, FIRST or NEXT action. Distractors are sensible actions that skip a step or belong to another task. Three items need a calculation or a table. Difficulty sits at or slightly above the live exam. All questions are written for this article, not taken from exam dumps.
| Sub-topic | Questions |
|---|---|
| Plan data collection (2.2.1) | 1, 2 |
| Determine the data sets (2.2.2) | 3, 4, 5, 6 |
| Collect data (2.2.3) | 7, 8 |
| Validate data (2.2.4) | 9 |
| Data Dictionary, ETL, Interface Analysis | 10, 11, 12 |
CBDA exam at a glance
| Fact | Detail |
|---|---|
| Format | 75 multiple-choice, scenario-based questions |
| Time | 2 hours |
| Source Data weight | 15% (Domain 2) |
| Exam fee | US$405; US$250 for IIBA members |
| Passing score | Not published by IIBA |
Last verified: October 2026
For a mixed warm-up across all domains, try our first set of CBDA mock questions.
Practice Questions
Question 1
Northgate Home, a furniture retailer, wants to measure how weekend promotions affect basket size. Raw sales sit in the central data warehouse. A regional team keeps a hand-cleaned, aggregated spreadsheet of the same sales, and a recent audit found mismatches in it.
Which source approach is BEST?
- A. Use the spreadsheet, because it is already in a usable form.
- B. Assess the spreadsheet's quality and favor a direct warehouse pull if it is doubtful.
- C. Merge both sources, because a larger combined volume will reduce error.
- D. Ask the regional team to re-certify the spreadsheet, then use it alone.
Question 2
Kestrel Freight is studying delivery delays after route changes. The research question needs daily telematics data to show short-term effects, but the vendor delivers a monthly batch.
What should the analyst do NEXT?
- A. Proceed with monthly data and record the gap as a limitation.
- B. Ask IT to switch the vendor feed to daily delivery right away.
- C. Rewrite the research question so monthly data can answer it.
- D. Assess the cost and effort of obtaining the data daily.
Question 3
Halden Mutual, an insurer, meets to approve sources for a claims-leakage analysis. Claims, finance and IT disagree about which systems feed which repositories and how often, and the discussion stalls.
What should the analyst prepare to BEST support the decision?
- A. A conceptual diagram of the sources, the flows between them, and feed frequency.
- B. A data dictionary with the definition and format of every claim field.
- C. A short survey asking each stakeholder which source they prefer.
- D. A stakeholder map ranking system owners by influence and interest.
Question 4
Corvane Bank has three research questions for a lending study. Questions 1 and 2 can use internal data. Question 3 needs external credit-bureau data:
| Item | Annual amount |
|---|---|
| Bureau data licence | US$210,000 |
| Estimated benefit of answering Question 3 | US$95,000 |
What should the analyst do?
- A. Recommend buying the data, because external data strengthens insights.
- B. Answer Question 3 with internal proxy data without noting the change.
- C. Present cost and benefit and recommend dropping or deferring Question 3.
- D. Postpone the whole study until the budget covers all three questions.
Question 5
Bramblefield Grocers has 120,000 loyalty members: 72,000 shop mostly in store and 48,000 mostly online. A price-sensitivity survey in the mobile app returned 2,400 responses, 1,920 of them from online shoppers.
What should the analyst conclude FIRST?
- A. Online shoppers are over-represented, so a sample matching the member split is needed.
- B. The sample covers 2% of members, so the shopper mix does not matter.
- C. Doubling responses to 4,800 through the same app will remove the error.
- D. The in-store segment should be dropped because it responded less.
Question 6
Lumen Stream, a video service, is investigating first-month cancellations. An engineer proposes adding a huge data set of thumbnail hover events because it is already captured. No stakeholder can link hover events to a business goal for this study.
Which five Vs consideration should the analyst raise FIRST?
- A. Volume
- B. Velocity
- C. Variety
- D. Value
Want more than 12? Create a free account on certificate.tips to unlock more CBDA practice questions, including the full 75-question CBDA exam simulator with exam and practice modes. certificate.tips is our product.
Question 7
Arlo Stays, a hotel group, has a 40-question satisfaction survey ready for 25,000 guests, embedded in online checkout. The sponsor wants to launch next week.
What should the analyst recommend NEXT?
- A. Send it to all 25,000 guests at once to maximize responses.
- B. Run it with a small group of guests first to test the approach.
- C. Cut it to 10 questions and launch without further testing.
- D. Have front-desk managers approve the wording, then launch.
Question 8
Vorsted Components sees two defect rates for one month: 2.1% in the quality system and 3.4% in a report built from ERP data. The ERP figure passes through two staging spreadsheets.
What should the analyst do FIRST?
- A. Report the average of both figures until the cause is known.
- B. Use the ERP figure, because the ERP is the system of record.
- C. Trace where each figure originates, how it is transformed, and where it is stored.
- D. Escalate to the data governance council for a ruling.
Question 9
Mediq Pharmacies requested an April prescriptions extract. The analyst's checks:
| Check | Result |
|---|---|
| Rows | 10,000 |
| Distinct prescription IDs | 9,820 |
| Dispense dates | 1 to 30 March |
| Nulls in required fields | 0 |
| Front-end vs database, 50 sampled orders | All match |
Which two quality characteristics does the extract FAIL? (Choose two.)
- A. Accuracy
- B. Uniqueness
- C. Completeness
- D. Consistency
- E. Timeliness
Question 10
Pelham Energy is combining billing and field-service data. Billing calls an account "active" if it was billed in the last 90 days; field service calls it "active" if a meter is connected. The counts differ by thousands.
What should the analyst do BEST?
- A. Expand the data dictionary with aliases, meanings and business rules for each definition.
- B. Adopt the billing definition everywhere, because billing owns revenue.
- C. Show both counts on a dashboard and let users choose.
- D. Rename the field in the field-service system.
Question 11
Skyhaven Air loads bookings into a warehouse that supports incremental loads. The data feeds a weekly operations report and a no-show model retrained each quarter.
Which loading approach is BEST?
- A. Incremental loads for both, because they run faster.
- B. Feed the model from the weekly report's output tables.
- C. Incremental for the model, full extraction for the report.
- D. Incremental for the report, full extraction for the model's training data.
Question 12
Quillwise, an online education company, finds 18% more learners in its analytics database than on the CRM screen. Learners edit profiles through a web form, and nobody has checked how edits are stored.
What should the analyst do FIRST?
- A. Remove duplicates by email address and continue.
- B. Survey learners about whether they hold several accounts.
- C. Analyze how the profile form's fields map to database records on edit.
- D. Accept the CRM figure, because the business already uses it.
Ready for the full set? Sign up and create a free account to keep practicing. The CBDA exam simulator has a full 75-question set in timed exam mode or practice mode with explanations.
Answer Key and Explanations
1. Correct answer: B. When data exists in several places, the analyst picks the best source at the right granularity. The Guide contrasts a central warehouse with a peripheral, already-manipulated source: a direct pull means more work but is justified when the secondary data's quality is questionable. Source: Guide to Business Data Analytics, 2.2.1 Plan Data Collection Why the others are wrong: A ignores known defects; C mixes doubtful data in; D still relies on a hand-massaged source.
2. Correct answer: D. When needed frequency exceeds current collection, the Guide calls for assessing the cost of obtaining data at the more regular interval. It ties frequency to whether data affects short- or long-term decisions. Source: Guide to Business Data Analytics, 2.2.1 Plan Data Collection Why the others are wrong: A defeats the question; B commits spend unassessed; C bends the question to the data.
3. Correct answer: A. Analysts create conceptual architecture diagrams of sources, flows and feed frequency to facilitate sourcing discussions and approvals. Data flow diagrams show required sources, dependencies and how data reaches the target repository. Source: Guide to Business Data Analytics, 2.2.2 Determine the Data Sets; 3.5 Data Flow Diagrams Why the others are wrong: B defines fields, not flows; C gathers opinions; D maps people, not data.
4. Correct answer: C. The licence exceeds the benefit by US$115,000 a year. Analysts weigh cost, value, timing, risk and feasibility, especially for third-party data, and the Guide notes some research questions are dropped when data is too expensive. Source: Guide to Business Data Analytics, 2.2.2 Determine the Data Sets Why the others are wrong: A ignores the numbers; B hides a scope change; D blocks two answerable questions.
5. Correct answer: A. Online shoppers are 40% of members, so a proportional sample of 2,400 would hold about 960; this one holds 1,920 (80%). Sampling should yield a representative subset, and 3.17 names the sample group and statistical sampling skills as keys to unbiased results. Source: Guide to Business Data Analytics, 2.2.2 Determine the Data Sets; 3.17 Survey and Questionnaire Why the others are wrong: B confuses size with representativeness; C doubles the bias; D reshapes the question around bias.
6. Correct answer: D. In the five Vs assessment, value means analytics must be driven by real business goals. A data set nobody can tie to a goal fails before size or speed matter. Source: Guide to Business Data Analytics, 2.2.2 Determine the Data Sets Why the others are wrong: A is size; B is generation speed; C is format and source mix.
7. Correct answer: B. The Guide recommends testing a collection approach on a small number of observations before large-scale collection; for a survey, that is a pilot with a small group. Question design strongly shapes the data, so a pilot catches flaws cheaply. Source: Guide to Business Data Analytics, 2.2.3 Collect Data; 3.17 Survey and Questionnaire Why the others are wrong: A risks a flawed full launch; C redesigns without evidence; D is a review, not a test.
8. Correct answer: C. Knowing where data comes from, what transformations it undergoes and where it is stored is data lineage, which the Guide uses to assess quality. The staging spreadsheets are the first place to look. Source: Guide to Business Data Analytics, 2.2.3 Collect Data Why the others are wrong: A blends in a likely error; B assumes two manual steps were clean; D escalates before analysis.
9. Correct answers: B and E. 10,000 rows but 9,820 distinct IDs means 180 duplicates, failing uniqueness. March dates for an April request fail timeliness, which the Guide checks by asking whether data covers the requested period. Source: Guide to Business Data Analytics, 2.2.4 Validate Data Why the others are wrong: A passes (totals match); C passes (no nulls); D shows no conflicting values.
10. Correct answer: A. A data dictionary standardizes definitions within and across sources. Its basic elements, such as name, aliases, meaning and description, can be extended to track business rules, which fits two valid definitions of one word. Source: Guide to Business Data Analytics, 3.4 Data Dictionary Why the others are wrong: B discards a valid view; C shows the conflict unresolved; D edits a source system to fix a definition.
11. Correct answer: D. Incremental loads add only changed data and speed up BI reports, while a full extraction that replaces target data supports better predictive and prescriptive analytics. Source: Guide to Business Data Analytics, 3.10 Extract, Transform, and Load (ETL) Why the others are wrong: A favors speed over model needs; B feeds a model report-shaped data; C reverses the load types.
12. Correct answer: C. Interface analysis links interface elements to database attributes and surfaces hidden business rules. The Guide's example is a profile update that either changes the record or creates a new one flagged as most recent, which inflates naive counts. Source: Guide to Business Data Analytics, 3.13 Interface Analysis Why the others are wrong: A treats a symptom; B asks users about invisible system behavior; D ignores a known gap.
For a deeper walkthrough of these tasks, read our guide to sourcing and validating data.
How to score this set
Count Question 9 as correct only if you chose both B and E.
- 11–12: you read Domain 2 the way the exam does; move to timed full-length practice.
- 8–10: reread the cited sections for your misses, especially where planning ends and collecting begins.
- Below 8: work through 2.2.1 to 2.2.4 again, then retake this set in a few days.
Twelve questions are a narrow sample. For broader coverage, run the full 75-question CBDA exam simulator, and if you have no account yet, sign up for free.
What candidates say about source data on the CBDA
Candidates who have passed describe a test that rewards judgement over recall. The author of our CBDA certification overview flagged about 20 of 75 questions and found the live exam harder than practice banks, with more weight on analytical decisions than definitions. Mufaddal Haidermota, who earned the CBDA in December 2022, advises per-domain mock questions, full-length tests and mind maps. A Techcanvass candidate story describes a developer who leaned on mind maps and suggests two to three hours of daily study. Trainer LN Mishra of Adaptive US recommends daily practice questions, timed mocks and reviewing every wrong answer, advice practitioners can apply directly to this domain.
FAQ
What does the Source Data domain cover on the CBDA?
It covers planning data collection, determining the data sets, collecting data and validating data. IIBA weights it at 15% of the 75-question exam. Questions usually ask which action comes first or which source or technique fits.
Are CBDA source data questions technical?
They expect you to understand lineage, sampling, ETL load types and quality checks, but not to write code. Most items test whether you choose the right step in the right order, with business meaning weighted as much as technology.
Which data quality characteristics should I know for Domain 2?
The Guide names accuracy, completeness, consistency, uniqueness and timeliness. Learn one example check for each, such as duplicate counts for uniqueness or the requested period for timeliness. Scenarios often describe the check rather than name the characteristic.
Sources
- Certification in Business Data Analytics (CBDA) — IIBA, checked October 2026
- Certification Fees — IIBA, checked October 2026
- CBDA Certification Handbook — IIBA, checked October 2026
- Business Data Analytics Certification — Mufaddal Haidermota, Medium, January 2023
- CBDA Certification Journey: Ivan — Techcanvass, 2024
- Mastering the CBDA Exam — LN Mishra, Adaptive US, October 2024