MAC Data Analysis & Decision Making 3 — Questions and Answers
Question 1: A dataset shows that ice cream sales and drowning rates both rise in summer. This is best described as:
- A causal relationship driven by ice cream consumption
- Spurious correlation caused by a confounding variable (hot weather) (Correct answer)
- A negative correlation between the two variables
- Statistical noise requiring more data collection
Correct answer: Spurious correlation caused by a confounding variable (hot weather)
Both variables are driven by the confounding variable of hot weather, making the correlation spurious rather than causal.
Question 2: Which statistical test is most appropriate for determining whether two independent group means are significantly different?
- Chi-square test
- Independent samples t-test (Correct answer)
- Linear regression
- Cluster analysis
Correct answer: Independent samples t-test
The independent samples t-test is used to compare the means of two unrelated groups to determine if the difference is statistically significant.
Question 3: A marketing team reports a 200% improvement in click-through rate. The absolute CTR changed from 0.1% to 0.3%. What should an analyst flag?
- The relative improvement is too small to matter
- The absolute values are low and context is needed to assess business impact (Correct answer)
- A 200% lift always indicates test fraud
- The CTR cannot increase by more than 100%
Correct answer: The absolute values are low and context is needed to assess business impact
Reporting only relative improvement can be misleading when the absolute baseline is very low, so both metrics must be contextualized.
Question 4: In multi-touch attribution, what distinguishes a data-driven model from a rules-based model such as last-touch?
- Data-driven models assign credit using machine learning on actual conversion path data (Correct answer)
- Rules-based models require more data to build than data-driven models
- Data-driven attribution always gives equal credit to every touchpoint
- Last-touch attribution is more accurate because it uses regression analysis
Correct answer: Data-driven models assign credit using machine learning on actual conversion path data
Data-driven attribution uses algorithms trained on observed conversion paths to assign fractional credit, rather than applying fixed heuristic rules.
Question 5: When building a customer lifetime value (CLV) model, which input is least relevant?
- Average order value
- Purchase frequency
- Customer's ZIP code demographics (Correct answer)
- Gross margin per transaction
Correct answer: Customer's ZIP code demographics
ZIP code demographics are not a direct input to CLV; the core inputs are purchase behavior and margin data.
Question 6: A marketing analyst uses a moving average to smooth a time series. What does this technique primarily help reveal?
- Individual data anomalies and outliers
- Underlying trends by reducing short-term noise (Correct answer)
- Seasonal patterns at their highest resolution
- The statistical significance of week-over-week changes
Correct answer: Underlying trends by reducing short-term noise
A moving average dampens short-term fluctuations to make longer-term trends more visible in time series data.
Question 7: What does a p-value of 0.03 indicate in the context of an A/B test?
- There is a 3% chance the result is correct
- There is a 97% probability that the variant is better than the control
- There is a 3% probability of observing this result (or more extreme) if the null hypothesis were true (Correct answer)
- The test needs 97% more data before concluding
Correct answer: There is a 3% probability of observing this result (or more extreme) if the null hypothesis were true
A p-value is the probability of observing results at least as extreme as those obtained, assuming the null hypothesis is true.
MAC Data Analysis & Decision Making 3
A dataset shows that ice cream sales and drowning rates both rise in summer.
This is best described as: