ISYE 6501 Final Quiz - Intro Analytics Modeling Questions and Answers (Latest Update 2026)

1



ISYE 6501 Final Quiz - Intro Analytics Modeling
Questions and Answers (Latest Update 2026)



Q1. When do you use hypothesis tests?
• A) When you have a small sample size
• B) When you know the underlying distribution of the data
• C) When you do not know the underlying distribution
• D) When you are comparing more than two groups

✅ Correct Answer: B
Rationale: Parametric hypothesis tests assume that the data follow a specific
distribution (e.g., normal). If the distribution is unknown or the assumptions are
violated, nonparametric tests are preferred.
Q2. What type of test should you conduct if you do not know the underlying
distribution of the data?
• A) t-test
• B) ANOVA
• C) Non-parametric test
• D) Chi-square test

✅ Correct Answer: C
Rationale: Non-parametric tests do not assume a particular distribution and are
used when distributional assumptions are violated or the underlying distribution
is unknown.
Q3. Nonparametric tests are useful when … (Select all that apply.)
• A) We don’t know much about the form of the underlying distribution

pg. 1

,2


• B) It’s important to have information about the median
• C) We have a very large dataset
• D) We don’t have much data

✅ Correct Answer: A, B, and D
Rationale: Nonparametric tests are valuable when the distribution is unknown,
when inference about the median is needed, or when the sample size is small.
Q4. McNemar’s test is used for:
• A) Continuous outcomes
• B) Paired categorical data
• C) Time series forecasting
• D) Clustering validation

✅ Correct Answer: B
Rationale: McNemar’s test is applied to paired nominal data (e.g., before-after
studies with binary outcomes). It evaluates whether the row and column marginal
frequencies are equal.
Q5. Which of the following best describes the Wilcoxon signed-rank test? (Select
all that apply.)
• A) Assumes the response function is continuous and symmetric
• B) Tests whether the median differs from a specified value
• C) Used for two independent samples
• D) Used for paired samples or one sample

✅ Correct Answer: A, B, and D
Rationale: The Wilcoxon signed-rank test is a nonparametric alternative to the
paired t-test. It assumes continuity and symmetry and tests the median against a
hypothesized value or the median difference between paired observations.




pg. 2

,3


Section 2: A/B Testing & Foundational Concepts
Q6. What is the primary purpose of A/B testing?
• A) To test three or more alternatives simultaneously
• B) To test two alternatives to see which one leads to better outcomes
• C) To validate model assumptions before data collection
• D) To determine the underlying distribution of data

✅ Correct Answer: B
Rationale: A/B testing is a controlled experiment that compares two versions (A
and B) to determine which performs better on a specific outcome measure.
Q7. Which of the following statements about bias and variance is true?
• A) As model complexity increases, bias increases.
• B) As model complexity increases, variance decreases.
• C) As model complexity increases, variance increases.
• D) Bias and variance are not affected by model complexity.

✅ Correct Answer: C
Rationale: In the bias-variance tradeoff, simple models have high bias
(underfitting) and low variance, while complex models have low bias but high
variance (overfitting).
Q8. Overfitting occurs when:
• A) The model performs well on test data
• B) The model is too simple
-C) The model captures noise in the training data
• D) Bias is high

✅ Correct Answer: C
Rationale: Overfitting happens when a model learns random noise or irrelevant
patterns in the training data, leading to poor generalization on unseen data.

pg. 3

, 4


Q9. The primary purpose of a train-test split is to:
• A) Increase accuracy artificially
• B) Evaluate model generalization to unseen data
• C) Reduce the number of features
• D) Normalize the data

✅ Correct Answer: B
Rationale: By training on one subset and testing on a separate held-out subset,
the train-test split estimates how well the model will perform on new, unseen
data.
Q10. K-fold cross-validation helps to:
• A) Reduce bias only
• B) Estimate model performance more reliably
• C) Increase the size of the dataset artificially
• D) Remove outliers

✅ Correct Answer: B
Rationale: K-fold cross-validation provides a more robust estimate of model
performance by averaging results over multiple splits, reducing the variability of
the performance metric.


Section 3: Regression & Classification
Q11. Which of the following is NOT an assumption of linear regression?
• A) Linearity
• B) Independence of errors
• C) Normality of predictors
• D) Homoscedasticity



pg. 4