Compare Two Categorical or Two Continuous Features Using the Appropriate Method MACHINE-LEARNING-ASSOCIATE Practice Questions (Page 1)

Questions 1–5

1. 1application · medium

   A data scientist wants to compare the mean CPU usage between two groups of servers using a t-test. The sample sizes are 15 and 12. A Shapiro-Wilk test on the first group returns p = 0.08 and on the second group p = 0.06. Levene's test returns p = 0.03. Which test should be used?

   • AWelch's t-test
   • BMann-Whitney U test
   • CStandard independent t-test
   • DPaired t-test
   Reveal Solution Discussion Select an answer first
2. 2application · medium

   A data analyst wants to compare the average transaction amount between two customer segments (new vs. returning). They have 200 transactions per segment. A Shapiro-Wilk test on each segment returns p-values of 0.001 and 0.002. What is the most appropriate test?

   • AWelch's t-test
   • BKruskal-Wallis test
   • CMann-Whitney U test
   • DIndependent t-test
   Reveal Solution Discussion Select an answer first
4. 3application · medium

   A data scientist wants to compare the average daily active users (DAU) between two app versions (v1 and v2). They have 30 days of DAU data for each version. A Shapiro-Wilk test on each sample returns p-values of 0.21 and 0.34, and Levene's test for equal variances returns p = 0.45. Which test is most appropriate?

   • AIndependent two-sample t-test
   • BMann-Whitney U test
   • CPaired t-test
   • DWelch's t-test
   Reveal Solution Discussion Select an answer first
5. 4foundation · easy

   A data scientist wants to determine whether there is a statistically significant association between two categorical features: customer segment (e.g., 'Retail', 'Wholesale', 'Enterprise') and product category (e.g., 'Hardware', 'Software'). Which statistical test is most appropriate for this comparison?

   • ATwo-sample t-test
   • BPearson correlation coefficient
   • CChi-square test of independence
   • DANOVA (F-test)
   Reveal Solution Discussion Select an answer first
6. 5foundation · easy

   A machine learning engineer is analyzing a 2x2 contingency table with very small expected cell counts (e.g., some cells have an expected count of 2). Which statistical test is more appropriate than the standard chi-square test for comparing these two categorical features?

   • AKruskal-Wallis test
   • BFisher's exact test
   • CPaired t-test
   • DMann-Whitney U test
   Reveal Solution Discussion Select an answer first