I85PREP
⌘ K
IIBA-CBDA study materials
IIBA-CBDA study resources

IIBA-CBDA practice questions

One question at a time. Build your IIBA-CBDA understanding.

First 50 of 200 questions. Each one lists the choices, the correct answer, and an explanation. The checker at the top of this page saves progress in this browser when it is available.

Continue in the practice tool above to load more questions, 50 at a time.

Question 1Identify the Research Questions

What is the primary purpose of a research question in a business data analytics initiative?

  1. To focus the analytics effort on a specific, answerable inquiry tied to a business problem Correct answer.
  2. To document the statistical models and algorithms that will be built once the data has been collected and cleaned
  3. To record the tooling and platform architecture decisions that were made during the early project planning phase
  4. To list every stakeholder who might be interested in the outcome of the analysis, in priority order

Explanation

A research question states the specific inquiry the analytics work must resolve, anchoring data sourcing, analysis, and reporting to a real business need. Model selection and tooling are downstream delivery decisions made only after the question is set.

Question 2Identify the Research Questions

Which characteristic most reliably distinguishes a well-formed research question?

  1. It is broad enough to accommodate any related topic that stakeholders raise later in the initiative
  2. It specifies a measurable outcome for a defined population or process and can be answered with data Correct answer.
  3. It deliberately avoids naming any metric so that analysts retain maximum flexibility
  4. It states the preferred solution that the business sponsor wants the analytics team to validate

Explanation

A usable research question names something measurable about a defined population or process, so that evidence can settle it. Breadth, solution preference, and deliberate vagueness each prevent the question from being answered or tested.

Question 3Identify the Research Questions

A retail sponsor reports that sales have fallen for two consecutive quarters and asks the analytics team to 'build a dashboard.' Which framing is the strongest research question?

  1. Which business intelligence platform should the retailer license to display quarterly sales data?
  2. How can the retailer grow its social media follower base among younger shoppers?
  3. Should the retailer replace the current sales leadership team?
  4. Which product, channel, and customer factors explain the sales decline over the past two quarters? Correct answer.

Explanation

A research question restates the business problem as a specific, answerable inquiry - here, which factors drive the decline. Requests for a tool or a personnel change presuppose a solution and foreclose examination of the cause.

Question 4Identify the Research Questions

In the SMART framework applied to research questions, what does the 'M' require?

  1. That the question receives mandatory sign-off from the project manager
  2. That the question can be automated in a reporting tool
  3. That the question identifies a measurable outcome or metric Correct answer.
  4. That the question references multiple variables at once

Explanation

In SMART, 'M' stands for measurable: the question must involve an outcome or metric that can be quantified. Variable count, approvals, and automation are not part of that criterion.

Question 5Identify the Research Questions

A team has documented 'customer churn is rising' as its business problem. How should the research question relate to that statement?

  1. It should stay deliberately generic so that any finding can be reported against it
  2. It should repeat the statement word for word so nothing is lost in translation
  3. It should turn the concern into a specific, measurable inquiry about churn drivers Correct answer.
  4. It should replace the business problem entirely once the analysis begins

Explanation

The business problem describes the concern, while the research question translates it into a specific, measurable inquiry that analytics can resolve. Both artefacts persist, and the question adds direction for data and analysis choices.

Question 6Identify the Research Questions

Within the FINER framework for evaluating research questions, which criterion asks whether the question can be answered with the available data, time, and skills?

  1. Ethical soundness
  2. Novelty
  3. Relevance
  4. Feasibility Correct answer.

Explanation

In FINER, feasibility asks whether the question can realistically be answered with the data, time, expertise, and resources available. Novelty, ethics, and relevance address different dimensions of quality.

Question 7Identify the Research Questions

Which research question is descriptive in nature?

  1. What proportion of policyholders lapsed in each of the last four quarters? Correct answer.
  2. What will renewal rates be next quarter if the discount ends?
  3. What segments of lapsed customers might exist that the business has not yet identified?
  4. Why does the new pricing tier increase renewals in one region but not another?

Explanation

Descriptive questions characterize what is happening - counts, proportions, and distributions - without explaining causes or forecasting. Causal, predictive, and exploratory questions serve other purposes.

Question 8Identify the Research Questions

Which research question is explanatory rather than predictive?

  1. How many customers opened the app last month by channel?
  2. How many customers are expected to open the app next month?
  3. Does the redesigned onboarding flow reduce time to first purchase? Correct answer.
  4. What groups of customers might exist based on app usage patterns?

Explanation

An explanatory question tests whether one factor produces an effect on another, as in whether a redesigned flow changes time to first purchase. Prediction, description, and exploration answer different kinds of questions.

Question 9Identify the Research Questions

Before finalizing a research question about 'high-value customers,' what should the analyst do first?

  1. Select the statistical test that will be used for the comparison
  2. Ask the sponsor to approve the final report format
  3. Begin collecting all available customer data to see what emerges
  4. Define 'high-value' with an explicit, measurable threshold that stakeholders agree on Correct answer.

Explanation

Ambiguous terms such as 'high-value' must be pinned down with an operational definition - an explicit, agreed, measurable threshold - before the question can be tested. Otherwise different stakeholders will answer different questions with the same data.

Question 10Identify the Research Questions

A sponsor wants to know why service quality varies across five business units over three years, but only six months of data exist for two units. What is the most appropriate action?

  1. Decline the request and close the initiative because the data are incomplete
  2. Narrow the question to the units and period the data can support, and confirm the revised scope with the sponsor Correct answer.
  3. Replace the analysis with the team's expert opinion about service quality
  4. Proceed with the original question and footnote the missing data in the final report

Explanation

Domain 1 includes assessing feasibility, so the analyst renegotiates scope to what the data can support and confirms it with the sponsor. Silent gaps, abandonment, or substituting opinion all undermine the evidence base.

Question 11Identify the Research Questions

Why is early stakeholder engagement important when defining research questions?

  1. It helps confirm that the question reflects a real business need and that someone owns the resulting decision Correct answer.
  2. It removes the need to document the business problem in writing
  3. It guarantees that the analysis will require no further review once complete
  4. It allows analysts to skip feasibility checks on data availability

Explanation

Stakeholders define the business need, own the decision the analysis informs, and can confirm that the question is worth answering. Their involvement does not remove documentation, feasibility checks, or later review.

Question 12Identify the Research Questions

Which research question is solution-biased rather than problem-focused?

  1. Which customer segments abandon identity verification most often?
  2. Why do a fifth of new accounts fail identity verification on their first attempt?
  3. Which identity-verification vendor should the company purchase this year? Correct answer.
  4. How does identity-verification completion time vary by channel?

Explanation

A solution-biased question starts from an assumed remedy, such as a vendor purchase, rather than from the business problem to be understood. Problem-focused questions ask why something happens, where it happens, or to whom.

Question 13Identify the Research Questions

Which statement is a testable hypothesis rather than a research question?

  1. Why did conversion rates fall in March?
  2. How should the marketing function be restructured next year?
  3. What is the conversion rate by acquisition channel for the last four quarters?
  4. Customers who receive a reminder email convert at a higher rate than those who do not. Correct answer.

Explanation

A hypothesis is a testable statement predicting a relationship between variables, which analysis can confirm or refute. Research questions pose an inquiry rather than assert an expected relationship.

Question 14Identify the Research Questions

Specifying that a research question concerns individual retail transactions rather than households is a matter of defining the:

  1. Sampling frame budget
  2. Data governance policy
  3. Unit of analysis Correct answer.
  4. Reporting cadence

Explanation

The unit of analysis is the entity about which the question is asked - transactions, customers, households, or sites - and it determines how data are aggregated and analysed. Sampling, governance, and reporting cadence are separate concerns.

Question 15Identify the Research Questions

Why should each research question be traceable to a documented business objective?

  1. So that the analytics team can avoid documenting assumptions and constraints
  2. Because traceability eliminates the need for a feasibility assessment
  3. So that the analysis results can be connected to a decision and its expected business value Correct answer.
  4. Because objectives automatically determine which statistical technique is valid

Explanation

Traceability connects the analysis to the decision and value it is meant to influence, which is how analytics work is prioritised and justified. It does not remove assumptions, feasibility work, or method selection.

Question 16Identify the Research Questions

What is the best reason to define, in advance, the criteria by which a research question will be considered answered?

  1. It prevents scope drift and sets shared expectations about what evidence will count Correct answer.
  2. It allows the team to skip the operational definition of key terms
  3. It guarantees the analysis will produce a statistically significant result
  4. It removes the need to validate findings with stakeholders later

Explanation

Agreeing in advance what evidence would count as an answer keeps the team from drifting and gives stakeholders a shared standard for completion. It does not guarantee significance or remove the need for definitions and validation.

Question 17Identify the Research Questions

Which condition most often makes a research question unanswerable through analytics?

  1. The question is sponsored by more than one business unit
  2. The question addresses a topic already studied by a competitor
  3. The key terms are undefined and the variables involved have no available or collectible data Correct answer.
  4. The question is reviewed by both the business and technology teams

Explanation

Answerability depends on whether the variables can be measured and the terms are precise enough to test. A question with undefined concepts or no obtainable data cannot be resolved by analytics.

Question 18Identify the Research Questions

A stakeholder asks, 'Should we launch product X?' Why should the analyst reframe this before proceeding?

  1. Because product decisions must always be tested with a controlled experiment
  2. Because yes/no questions can never be informed by data analysis
  3. Because it presumes a binary decision and gives no measurable outcome, population, or timeframe to investigate Correct answer.
  4. Because launch decisions belong to finance rather than to analytics

Explanation

A binary 'should we' question offers no measurable outcome, population, or timeframe, so no analysis can settle it. It is better reframed around expected effects, conditions, or drivers that data can address.

Question 19Identify the Research Questions

Why are research questions documented with identifiers and linked to the business problem statement?

  1. To replace the need for a final report at the end of the analysis
  2. To reduce the number of stakeholders consulted during the initiative
  3. To satisfy the formatting requirements of the analytics tool being used
  4. To maintain traceability so that later data, analysis, and findings can be tied back to the question Correct answer.

Explanation

Identifiers and links to the business problem preserve traceability, so later data, analysis, and findings can be shown to answer the intended question. Documentation is not about tooling, stakeholder reduction, or replacing reporting.

Question 20Identify the Research Questions

Two executives want the analytics team to answer different questions using the same limited data. What is the most appropriate way to resolve this while defining research questions?

  1. Answer the question raised by the more senior executive without discussion
  2. Combine the two questions into one broad statement that covers both topics
  3. Pursue both questions in full and let the analysis timeline expand accordingly
  4. Facilitate a prioritization discussion linking each question to business value, urgency, and the decision it informs Correct answer.

Explanation

Competing questions are resolved through facilitated prioritisation that weighs business value, urgency, and the decision each would inform. This keeps scope realistic and preserves stakeholder buy-in.

Question 21Identify the Research Questions

A team proposes 'Is our customer data complete?' as its research question. Why is this weak for the Identify the Research Questions domain?

  1. It duplicates the operational definitions already captured in the glossary
  2. It addresses a data readiness prerequisite rather than a business inquiry that informs a decision Correct answer.
  3. It cannot be answered by any technique available to analysts
  4. It is too narrow to require any stakeholder involvement

Explanation

Whether the data are complete is a data readiness question that must be satisfied before analysis, not a business inquiry that informs a decision. Research questions in this domain should be anchored in a business problem.

Question 22Identify the Research Questions

A proposed research question can only be answered using individual customers' personal data. How should the analyst proceed?

  1. Replace the question with one about the analytics team's own processes
  2. Discard the question because personal data can never inform analytics
  3. Confirm that the use of the data is lawful and ethical, and adjust the question or data sources if it is not Correct answer.
  4. Answer the question using the personal data, since the business has access to it

Explanation

Research questions must be lawful and ethical, so the analyst confirms the basis for using personal data and adjusts the question or sources if needed. Proceeding on access alone, or abandoning the business need, are both inappropriate.

Question 23Identify the Research Questions

Which combination of elements makes a research question measurable?

  1. A budget, a project charter, and a risk log
  2. A sponsor name, a delivery date, and a reporting tool
  3. A metric, a defined population, and a timeframe Correct answer.
  4. A hypothesis, a p-value, and a sample size

Explanation

Measurability requires a metric, a defined population, and a timeframe, so that a result can be compared against an expectation. Sponsor names, hypotheses, and budgets do not make a question measurable.

Question 24Identify the Research Questions

A discovery session produces nine candidate research questions. Which approach best narrows them?

  1. Score each against business value, feasibility, and the decision it supports, then agree priorities with stakeholders Correct answer.
  2. Select the three that are easiest to compute with existing reports
  3. Keep all nine so that no stakeholder is disappointed
  4. Let the analytics team choose the ones that interest them most

Explanation

Candidate questions are narrowed using explicit criteria - typically business value, feasibility, and the decision supported - agreed with stakeholders. Keeping everything or choosing on convenience or interest alone produces unfocused work.

Question 25Identify the Research Questions

A sponsor states, 'Our problem is that the website is too slow.' How should the analyst treat this when framing the research question?

  1. Reject it because sponsors cannot identify their own problems
  2. Adopt it verbatim as the research question since it comes from the sponsor
  3. Treat it as a perceived cause and check whether the underlying business problem, such as lost conversion or revenue, is the real target Correct answer.
  4. Reframe it as a question about server response times only

Explanation

A sponsor's stated cause is often a symptom or an assumption, so the analyst validates whether the real business problem is being targeted. Taking it verbatim or dismissing it both fail to establish a sound research question.

Question 26Identify the Research Questions

In an analytics initiative, what is the primary purpose of defining research questions during business analysis?

  1. To document the technical architecture and deployment plan for the reporting solution
  2. To specify the exact data fields and tables that will be queried
  3. To estimate the project budget and resource allocation
  4. To translate the business problem into focused, answerable inquiries that guide the analysis Correct answer.

Explanation

Research questions bridge the business problem and the analysis by stating precisely what must be investigated. They give direction to data collection, method selection, and interpretation without prescribing technical implementation.

Question 27Identify the Research Questions

Which of the following best exemplifies a research question rather than a data request?

  1. What factors are linked to churn among small business accounts? Correct answer.
  2. Join the CRM and billing tables on customer ID.
  3. Extract all sales transactions from the last four quarters.
  4. Provide a list of customers who churned in 2023.

Explanation

A research question states an inquiry whose answer must be discovered through analysis. Data requests and transformation instructions describe inputs or operations rather than the knowledge being sought.

Question 28Identify the Research Questions

A team is framing a question to explore a new market where little prior knowledge exists. Which question form is most appropriate at this stage?

  1. A question that presumes a specific causal mechanism
  2. A broad open-ended question that allows themes to emerge Correct answer.
  3. A closed yes/no question with a predefined answer set
  4. A question limited to a single numeric metric

Explanation

When prior knowledge is limited, open-ended questions allow relevant variables and themes to be discovered before narrowing. Closed or metric-bound questions presuppose what matters.

Question 29Identify the Research Questions

A well-formed research question should explicitly name which element?

  1. The name of the executive sponsoring the work
  2. The software tool that will be used for analysis
  3. The statistical technique that will be applied
  4. The specific construct or phenomenon being investigated Correct answer.

Explanation

A research question must identify the construct or phenomenon under study so that variables, population, and measures can be derived. Tools and techniques are chosen later to serve that construct.

Question 30Identify the Research Questions

A business goal states 'improve customer retention.' Which revision best converts it into a focused research question?

  1. What is the churn rate among mid-market accounts over the last 12 months? Correct answer.
  2. How can we improve customer retention across every segment and region in the business?
  3. Why is retention important?
  4. Should the company invest in a new CRM platform to support retention?

Explanation

A focused research question narrows a broad goal to a defined population, measure, and time frame that can be answered with data. Broad or rhetorical restatements do not create a testable inquiry.

Question 31Identify the Research Questions

Which research question is compound and should be split before analysis?

  1. What is the average order value for all online purchases made during 2023 by region?
  2. What proportion of invoices were paid late in 2023?
  3. How many support tickets were escalated last quarter?
  4. Do pricing changes and promotions affect sales, and which segments respond most? Correct answer.

Explanation

A compound question joins two distinct inquiries with a conjunction, making a single clear answer impossible. Each component should be separated so it can be analyzed and validated independently.

Question 32Identify the Research Questions

A research question asks whether employee turnover has worsened. Which addition makes the question sufficiently time-bound?

  1. Listing the names of department managers
  2. Naming the HR system that stores turnover records
  3. Specifying the comparison periods and the employee population Correct answer.
  4. Selecting a visualization tool for the results

Explanation

Terms such as 'worsened' require an explicit baseline and comparison period to be answerable. Specifying the periods and the population being compared gives the question a measurable time frame.

Question 33Identify the Research Questions

An analyst wants to compare satisfaction between two service channels. Which question is best framed for that comparison?

  1. How does satisfaction differ between the phone and chat channels this year? Correct answer.
  2. What is the total number of customer complaints received?
  3. Is customer satisfaction generally good across the organization and all of its service channels?
  4. Which service channel employs the most staff?

Explanation

A comparative question names the groups being compared, the outcome of interest, and the period. Vague or single-group questions cannot support a valid comparison.

Question 34Identify the Research Questions

Which question is evaluative, asking how well something performs against a standard?

  1. To what extent does the new onboarding process meet its target completion rate? Correct answer.
  2. What is the monthly count of newly opened customer accounts across all regions?
  3. Which region generated the highest revenue last year?
  4. What are the demographic attributes of churned customers?

Explanation

An evaluative question compares observed performance with a defined target or criterion. Counting, ranking, and profiling questions describe rather than evaluate against a standard.

Question 35Identify the Research Questions

Which research question is most actionable for a decision maker?

  1. What is the academic meaning of customer loyalty and how has it been defined historically?
  2. Which of two proposed discount structures yields higher repeat purchase within six months? Correct answer.
  3. How many words appear in the company's mission statement?
  4. What color scheme is used on the current website?

Explanation

An actionable question is tied to a decision and produces evidence that can distinguish among options. Questions with no bearing on a choice may be interesting but do not support action.

Question 36Identify the Research Questions

A research question asks, 'Why are our newer employees less productive than tenured staff?' Which assumption should the analyst surface before proceeding?

  1. That newer employees are in fact less productive than tenured staff Correct answer.
  2. That productivity is unrelated to employee tenure
  3. That tenured staff will leave the organization
  4. That productivity cannot be measured

Explanation

Questions often embed a premise that has not been verified, and analysis built on a false premise wastes effort. The analyst should confirm the observed difference exists before investigating its cause.

Question 37Identify the Research Questions

A research question about supply chain delays is being scoped. What does defining its boundaries accomplish?

  1. It determines the statistical significance threshold automatically.
  2. It establishes which regions, time periods, and product lines are included or excluded. Correct answer.
  3. It guarantees that the analysis will produce a definitive causal explanation of every delay.
  4. It replaces the need for stakeholder review.

Explanation

Boundaries make explicit what the question covers and what it does not, preventing scope creep and mismatched expectations. They define inclusion criteria rather than guaranteeing results or replacing governance.

Question 38Identify the Research Questions

A stakeholder asks, 'Does our new training program cause higher sales?' What does this wording require the analyst to consider?

  1. Only a descriptive summary of total sales before and after the training program was introduced
  2. A simple count of employees who completed training
  3. A design that can support a causal claim, such as a comparison group or controlled experiment Correct answer.
  4. A review of the training vendor's marketing materials

Explanation

Causal language requires a design that rules out plausible alternative explanations, typically through comparison or control. A before-and-after summary alone cannot establish that training produced the change.

Question 39Identify the Research Questions

Which question is prescriptive rather than predictive?

  1. Which inventory allocation policy should we adopt to minimize stockouts? Correct answer.
  2. How many units were sold last quarter?
  3. Which customers are likely to churn next month?
  4. What will next quarter's demand be for each product line?

Explanation

Prescriptive questions ask which action or policy should be chosen to achieve an objective. Predictive questions estimate future outcomes, and descriptive questions summarize what has occurred.

Question 40Identify the Research Questions

After a stakeholder review, a research question is found to be too broad to answer within the project timeline. What is the most appropriate next step?

  1. Escalate to the sponsor to remove the timeline constraint.
  2. Rewrite the question to match whatever data is easiest to obtain.
  3. Abandon the question entirely and proceed directly with data collection using whatever sources are available.
  4. Split or narrow the question into scoped sub-questions that remain linked to the business problem. Correct answer.

Explanation

Refinement should preserve the link to the business problem while reducing the question to a scope that can be answered. Splitting into sub-questions keeps the inquiry traceable and feasible.

Question 41Source Data

Which of the following is an example of structured data?

  1. A folder of scanned supplier invoices stored as images
  2. A social media feed of free-text comments about a brand
  3. A relational table of orders with defined columns Correct answer.
  4. A recorded customer service call in an audio file

Explanation

Structured data conforms to a predefined model of typed fields arranged in rows and columns, which makes it directly queryable and comparable across records. The other items lack a predefined field structure and require interpretation before analysis.

Question 42Source Data

What is the primary purpose of data profiling when assessing a new source?

  1. To reveal patterns, anomalies, and quality issues in the data Correct answer.
  2. To document the business rules governing report distribution
  3. To convert source values into the target schema
  4. To encrypt sensitive columns before the extraction step

Explanation

Profiling examines the actual values in a source, such as null rates, distinct counts, ranges, and formats, to establish the condition of the data and the risks of using it. It is an assessment activity that precedes transformation and loading decisions.

Question 43Source Data

An analyst collects customer feedback by designing and running a survey specifically for the current initiative. How should this data be classified?

  1. Secondary internal data
  2. Reference data
  3. Derived data
  4. Primary data Correct answer.

Explanation

Primary data is collected first-hand for the specific question at hand, which is exactly what a purpose-built survey produces. Secondary data, by contrast, already exists because someone collected it for a different purpose.

Question 44Source Data

Checking how many required customer records have a populated postal code measures which data quality dimension?

  1. Consistency
  2. Uniqueness
  3. Completeness Correct answer.
  4. Timeliness

Explanation

Completeness concerns whether the values that are expected to be present actually exist in the data set. Counting populated values against the required fields is the standard way to measure it.

Question 45Source Data

Which of the following is metadata rather than data?

  1. The definition and permitted values of a product status field Correct answer.
  2. The total revenue generated by each product line
  3. The customer name recorded on an order record
  4. The status values assigned to individual product records

Explanation

Metadata describes data: what a field means, what values it may hold, where it comes from, and how it is governed. The other options are the content values themselves.

Question 46Source Data

Every twentieth invoice in a sequentially numbered list is selected for review. Which sampling technique does this describe?

  1. Systematic sampling Correct answer.
  2. Stratified sampling
  3. Cluster sampling
  4. Convenience sampling

Explanation

Systematic sampling selects items at a fixed interval through an ordered population, which is what choosing every twentieth record achieves. It is straightforward to apply when a reliable ordering exists.

Question 47Source Data

Which source provides unstructured data?

  1. A table of employee badge numbers and departments
  2. A collection of recorded service calls and email bodies Correct answer.
  3. A CSV extract of monthly sales totals
  4. A spreadsheet of supplier payment terms

Explanation

Recorded speech and free-form email text have no predefined field structure and must be interpreted before they can be analysed quantitatively. Tabular extracts and spreadsheets are structured by contrast.

Question 48Source Data

What does data lineage describe?

  1. The legal basis for processing personal information
  2. The order in which reports are refreshed for users
  3. The path data takes from origin through transformation to use Correct answer.
  4. The statistical distribution of values in each column

Explanation

Lineage records where data came from, what was done to it, and where it ended up, supporting impact analysis and traceability of reported results. It is a record of movement and change rather than of value distribution or permissions.

Question 49Source Data

A team needs a consolidated view of the data sources available to an initiative. Which artifact is most appropriate?

  1. A data source inventory listing each source, its owner, and its content Correct answer.
  2. A risk register recording project issues and mitigations
  3. A schedule of extraction tasks and their run times
  4. A glossary of the acronyms used in reporting

Explanation

A source inventory is the catalogue that records what sources exist, who owns them, what they contain, and how they can be accessed. It gives the team a single reference point for planning acquisition.

Question 50Source Data

An analyst requires daily sales figures, but the only sales system stores monthly aggregates. What is the correct assessment of this source?

  1. The source records are not unique across the period
  2. The source volume is too high for the required analysis
  3. The source has adequate timeliness but poor completeness
  4. The source granularity is too coarse to meet the requirement Correct answer.

Explanation

Granularity is the level of detail at which data is held. Monthly aggregates cannot be decomposed into daily figures, so the source cannot satisfy a daily requirement regardless of its other qualities.