Statistics CERTIFIED-DATA-SCIENCE-PRACTITIONER Practice Questions (Page 1)

Questions 1–5

1. 1foundation · medium

   Why is it important to report effect size in addition to p-values?

   • AEffect size indicates the probability of a Type II error
   • BEffect size determines whether the null hypothesis is true
   • CEffect size is required to compute the p-value
   • DEffect size helps assess the practical importance of a statistically significant result
   Reveal Solution Discussion Select an answer first
2. 2application · medium

   A data scientist compares two machine learning models on a test set of 1,000 samples. Model A achieves 85% accuracy and Model B achieves 84.5%. A McNemar test yields a p-value of 0.03. The scientist wants to recommend Model A. What additional information is most important to include?

   • AThe training time of each model to ensure Model A is faster.
   • BThe p-value from a different test, such as a t-test, to confirm the result.
   • CThe p-value alone is sufficient because it shows Model A is statistically significantly better.
   • DThe effect size (e.g., the difference in accuracy or a measure like Cohen's h) to show whether 0.5% is practically meaningful.
   Reveal Solution Discussion Select an answer first
4. 3foundation · medium

   What does the assumption of homoscedasticity mean in the context of regression analysis?

   • AThe independent variables are not correlated with each other
   • BThe residuals have constant variance across all levels of the independent variable
   • CThe residuals are normally distributed
   • DThe relationship between variables is linear
   Reveal Solution Discussion Select an answer first
5. 4expert · medium

   A data scientist is comparing the median time-to-completion for two project management tools. The data for Tool A (n=25) are normally distributed, but Tool B (n=25) has several extreme outliers. The scientist wants to compare the central tendency. Which approach is most appropriate?

   • ARemove the outliers from Tool B and then run a t-test.
   • BUse a two-sample t-test because Tool A is normal and the sample sizes are equal.
   • CUse the Mann-Whitney U test, which is robust to outliers and does not require normality.
   • DUse a paired t-test because both tools are used by the same team.
   Reveal Solution Discussion Select an answer first
6. 5application · medium

   A data scientist runs a one-way ANOVA to compare the mean response time of three server configurations. The Levene's test for homogeneity of variances returns a p-value of 0.02. What should the scientist do next?

   • AProceed with the ANOVA results because Levene's test is not important.
   • BRun a Kruskal-Wallis test only if the data are also non-normal.
   • CRemove the group with the largest variance to restore homogeneity.
   • DUse Welch's ANOVA, which does not assume equal variances.
   Reveal Solution Discussion Select an answer first