Free IIBA-CBDA sample questions

Correct answer: These points represent potential outliers. The next step is to investigate their context and validity to determine if they are legitimate values or data entry errors.

A box plot is a standard tool for identifying potential outliers, which are data points that fall outside the typical range (often defined as 1.5 times the interquartile range above the third quartile or below the first). The correct first step is never to blindly remove them. An analyst must investigate these points to understand their cause. They could be high-value customers, data entry mistakes, or represent a unique but valid segment. Investigation precedes any action like removal, transformation, or imputation.

Correct answer: Which donor demographic and behavioral characteristics are most strongly correlated with high donation frequency and value?

This is the most effective research question because it is specific, data-driven, and directly addresses the business need. It moves beyond the simple hypothesis ('younger donors give less') to identify the actual characteristics of the most valuable donors. The results will allow the non-profit to create targeted, efficient campaigns instead of relying on assumptions.

Correct answer: A failure to translate technical model outputs into a meaningful business context for the audience.

The core issue is a communication failure. The team presented technical metrics ('feature importance') without explaining what they mean in a business context. The manager's distrust stems from the model being a 'black box' that contradicts their domain expertise ('my team's current pipeline'). Effective communication involves translating technical results into a narrative that resonates with the audience's experience and helps them understand how the model arrives at its conclusions.

Correct answer: It directly links an analytical insight (bundling reduces churn) to a specific, measurable, actionable, relevant, and time-bound (SMART) business goal.

This recommendation is effective because it perfectly translates a statistical finding into a concrete business strategy. It follows the SMART criteria: Specific (increase bundled policies), Measurable (15%), Actionable (offer a discount), Relevant (to the insight about churn), and Time-bound (next fiscal year). This creates a clear path from data insight to business action and measurable outcome.

Correct answer: Conduct an analytics maturity assessment to baseline the current capabilities, tools, skills, and processes across the organization.

Before creating a strategy or selecting tools, it is crucial to understand the current state. An analytics maturity assessment evaluates the organization's existing people, processes, and technology. This provides a clear baseline, identifies inconsistencies and gaps (like data silos), and informs the development of a realistic and targeted strategic roadmap. Jumping to a tool selection or governance model without this understanding often leads to failure.

Correct answer: Replace the missing values with the mean or median of the 'Years_of_Experience' column.

With 20% of values missing, deleting the rows (listwise deletion) would discard a significant portion of the dataset and potentially introduce bias. Replacing missing values with a measure of central tendency like the mean (for symmetric distributions) or median (for skewed distributions) is a standard and reasonable first approach. This technique, called imputation, preserves the sample size without drastically distorting the feature's overall distribution. More advanced techniques exist, but mean/median imputation is a common and appropriate starting point.

Correct answer: False

This statement is false. A high R-squared value simply means the model explains a large proportion of the variance in the training data. However, it can be misleading. A model can be overfit, meaning it has learned the noise in the training data too well. Such a model will have a high R-squared on the data it was trained on but will fail to generalize and perform poorly on new, unseen data. Other diagnostics, such as checking residual plots and using an adjusted R-squared, are necessary to assess model reliability.

Correct answer: CATWOE (Customers, Actors, Transformation, Weltanschauung, Owners, Environmental constraints)

CATWOE is specifically designed to analyze a problem situation from multiple stakeholder perspectives. It would compel the analyst to identify the Customers (shoppers), Actors (inventory managers, buyers), the Transformation process (raw materials to sold goods), the Weltanschauung (the worldview that inventory optimization is critical), the Owners (company executives), and Environmental constraints (fast-fashion trends, supply chain issues). This comprehensive view is ideal for framing a complex business problem like inventory optimization.

Correct answer: The analysis confuses correlation with causation, likely due to a confounding variable.

This is a classic example of the principle 'correlation does not imply causation'. It is highly likely that a third, confounding variable—such as sunny, hot weather—is causing an increase in both ice cream sales and sunglasses sales independently. Acting on the assumption that one causes the other (causation) could lead to an ineffective marketing strategy. The correct interpretation is that the two are related, but one does not necessarily cause the other.

Correct answer: Incorporate qualitative business context into the interpretation of quantitative data.

This scenario is a critical lesson in business data analytics: quantitative data (the 'what', 0.5% usage) is incomplete without qualitative context (the 'who' and 'why', valuable enterprise clients). A purely data-driven decision based on the usage metric would have been disastrous for the business. The product manager's domain knowledge provided the essential context that the raw numbers lacked. Effective decision-making requires synthesizing both types of information.