Download Statistical Inference: Confidence Intervals and Hypothesis Tests and more Exams Statistics in PDF only on Docsity!
Statistical Inference: Confidence Intervals and Hypothesis Tests Course: Data Analysis & Statistical Methods Total Questions: 150 Question Type: Multiple Choice Description: This comprehensive examination covers the core principles of statistical inference, focusing on confidence interval estimation and hypothesis testing. Topics include the logic of inference, sampling distributions and the Central Limit Theorem, standard error, z-intervals and t-intervals for population means, confidence intervals for proportions, determining sample size, the formulation of null and alternative hypotheses, Type I and Type II errors, significance levels, test statistics, p-value interpretation, and the relationship between confidence intervals and hypothesis tests. Emphasis is placed on proper interpretation, underlying assumptions, and practical applications in data analysis. This exam is suitable for undergraduate students in statistics, data science, social sciences, and business analytics.
1. Statistical inference is primarily concerned with: A) Organizing data into tables and charts B) Drawing conclusions about a population based on a sample C) Collecting data from every member of a population D) Describing the sample using numerical measures Answer: B Explanation: Statistical inference uses sample data to make generalizations or draw conclusions about the larger population from which the sample was drawn. 2. A numerical summary of a population is called a: A) Statistic B) Parameter C) Sample D) Variable Answer: B Explanation: A parameter is a numerical characteristic of a population (e.g.,
population mean μ, population proportion p). A statistic is the corresponding measure calculated from a sample.
3. A numerical summary of a sample is called a: A) Parameter B) Population C) Statistic D) Constant Answer: C Explanation: A statistic is a value computed from sample data (e.g., sample mean x/, sample proportion p0) used to estimate an unknown population parameter. 4. The standard deviation of the sampling distribution of a statistic is called the: A) Variance B) Standard error C) Parameter error D) Sample deviation Answer: B Explanation: The standard error measures the variability of a sample statistic across repeated samples from the same population. 5. According to the Central Limit Theorem, for a sufficiently large sample size, the sampling distribution of the sample mean will be approximately: A) Uniform B) Normal, regardless of the population distribution C) Skewed, matching the population skew D) Binomial Answer: B Explanation: The Central Limit Theorem states that the sampling distribution of x/ approaches a normal distribution as n increases, regardless of the shape of the original population distribution. 6. The Central Limit Theorem applies to the sample mean when the sample size is generally considered to be large, typically: A) n ≥ 5 B) n ≥ 15 C) n ≥ 30 D) n ≥ 100 Answer: C Explanation: A sample size of n ≥ 30 is often used as a rule of thumb for
11. The general form of a confidence interval is: A) Point estimate ± Standard error B) Point estimate ± Margin of error C) Point estimate × Confidence level D) Standard error ± Confidence level Answer: B Explanation: Every confidence interval is constructed as the point estimate plus and minus a margin of error. 12. The margin of error in a confidence interval is calculated as: A) Critical value × Standard error B) Critical value × Sample size C) Standard error × Sample mean D) Confidence level × Standard error Answer: A Explanation: Margin of error is the product of a critical value (z* or t*) and the standard error of the estimator. 13. If we increase the confidence level from 90% to 95%, the margin of error will: A) Increase B) Decrease C) Remain the same D) Double Answer: A Explanation: Higher confidence requires a larger critical value, which increases the margin of error and widens the interval to provide more coverage certainty. 14. If we increase the sample size, the width of the confidence interval will: A) Increase B) Decrease C) Remain the same D) Become unpredictable Answer: B Explanation: Increasing n decreases the standard error (σ/√n), thereby reducing the margin of error and producing a narrower interval. 15. The correct interpretation of a 95% confidence interval (a, b) for a population mean μ is: A) There is a 95% probability that μ is between a and b B) We are 95% confident that μ falls between a and b
C) 95% of the data falls between a and b D) μ will fall between a and b exactly 95% of the time Answer: B Explanation: Once computed, the interval either contains μ or not. The 95% refers to the long-run proportion of intervals from repeated sampling that would contain μ.
16. Which critical value (z) is used for a 90% confidence interval under the standard normal distribution? A) 1. B) 1. C) 2. D) 1. Answer: A Explanation: The z value for 90% confidence leaves 5% in each tail, corresponding to z = 1.645. 17. Which critical value (z) is used for a 95% confidence interval under the standard normal distribution? A) 1. B) 1. C) 2. D) 1. Answer: B Explanation: The z value for 95% confidence leaves 2.5% in each tail, corresponding to z = 1.96. 18. Which critical value (z) is used for a 99% confidence interval under the standard normal distribution? A) 1. B) 1. C) 2. D) 2. Answer: C Explanation: The z value for 99% confidence leaves 0.5% in each tail, corresponding to z = 2.576. 19. When constructing a confidence interval for a population mean and the population standard deviation σ is known, we use the: A) t-distribution B) z-distribution C) F-distribution D) Chi-square distribution
24. A 95% confidence interval for μ is (22.4, 28.6). What is the margin of error? A) 3. B) 6. C) 25. D) 4. Answer: A Explanation: Margin of error = (Upper - Lower) / 2 = (28.6 - 22.4) / 2 = 6.2 / 2 = 3.1. 25. Which of the following would produce the narrowest confidence interval for μ, all else being equal? A) 90% confidence with n = 100 B) 95% confidence with n = 50 C) 99% confidence with n = 200 D) 90% confidence with n = 30 Answer: A Explanation: A 90% confidence level gives a smaller critical value, and n=100 gives a smaller standard error than n=30, resulting in the narrowest interval among the options. 26. Which factor does NOT affect the width of a confidence interval for the mean? A) Sample size B) Population mean C) Confidence level D) Sample standard deviation Answer: B Explanation: The population mean is the target parameter being estimated; it does not influence the width of the interval. 27. To estimate the population proportion p, the standard error of the sample proportion p0 is: A) √(p(1-p)/n) B) √(p0(1-p0)/n) C) p0(1-p0)/n D) √(p0 / n) Answer: B Explanation: Since p is unknown, we estimate the standard error using the sample proportion: SE(p0) = √(p0(1-p0)/n). 28. The confidence interval for a population proportion p is calculated as: A) p0 ± z* × √(p0(1-p0)/n)
B) p0 ± t* × √(p0(1-p0)/n) C) x/ ± z* × (σ/√n) D) p0 ± z* × (p0/√n) Answer: A Explanation: The large-sample confidence interval for a proportion uses the z-distribution and the estimated standard error of p0.
29. A sample of 100 students shows that 60% prefer online learning. The 95% confidence interval for the true proportion is approximately: A) (0.50, 0.70) B) (0.56, 0.64) C) (0.40, 0.80) D) (0.55, 0.65) Answer: A Explanation: SE = √(0.6×0.4/100) = √0.0024 ≈ 0.049. Margin of error = 1.96×0.049 ≈ 0.096. Interval = 0.60 ± 0.096 ≈ (0.504, 0.696), which rounds to (0.50, 0.70). 30. The formula to determine the sample size needed for a desired margin of error E when estimating μ (with known σ) is:
A) n = (z σ / E)²
B) n = (zE / σ)²
C) n = σ² / (z E)
D) n = (zσ) / E
Answer: A Explanation: Solving E = z* × (σ/√n) for n yields n = (z*σ / E)².
31. In hypothesis testing, the null hypothesis (H₀) is a statement of: A) No effect or no difference B) The effect we hope to prove C) The alternative claim D) The sample result Answer: A Explanation: The null hypothesis typically represents the status quo or a statement of no effect, which is tested against an alternative hypothesis. 32. The alternative hypothesis (Hₐ) is a statement that: A) Is assumed true unless evidence suggests otherwise B) Will be accepted if the null hypothesis is accepted C) Contradicts the null hypothesis and is what the researcher wants to support D) Is always two-sided
C) Reject a false null hypothesis D) Fail to reject a true null hypothesis Answer: B Explanation: Type II error (β) occurs when we fail to reject H₀ even though it is false.
38. The power of a hypothesis test is defined as: A) The probability of rejecting a true null hypothesis B) The probability of failing to reject a false null hypothesis C) The probability of rejecting a false null hypothesis D) The significance level Answer: C Explanation: Power = 1 - β, and it is the probability of correctly detecting an effect when one truly exists (i.e., rejecting H₀ when it is false). 39. If we decrease α (e.g., from 0.05 to 0.01), the power of the test will generally: A) Increase B) Decrease C) Remain the same D) Become 1 Answer: B Explanation: Lowering α makes the rejection region smaller, making it harder to reject H₀, thus reducing the probability of rejecting a false H₀ (power). 40. Increasing the sample size will generally: A) Decrease power B) Have no effect on power C) Increase power D) Increase α Answer: C Explanation: Larger sample sizes reduce standard error, making test statistics more sensitive to departures from H₀, thereby increasing the power of the test. 41. A test statistic is: A) The critical value used to define the rejection region B) A value calculated from the sample data that is used to decide whether to reject H₀ C) The value of the population parameter D) The probability of making a Type I error
Answer: B Explanation: A test statistic standardizes the sample estimate relative to the null hypothesis value, allowing comparison to a known distribution.
42. The z-test statistic for a population mean when σ is known is calculated as: A) (x/ - μ₀) / (σ / √n) B) (x/ - μ₀) / (s / √n) C) (x/ - μ₀) / σ D) (p0 - p₀) / √(p₀(1-p₀)/n) Answer: A Explanation: When σ is known, the z statistic standardizes the difference between the sample mean and the hypothesized population mean by the standard error σ/√n. 43. The t-test statistic for a population mean when σ is unknown is calculated as: A) (x/ - μ₀) / (σ / √n) B) (x/ - μ₀) / (s / √n) C) (x/ - μ₀) / s D) (x/ - μ₀) × (√n / σ) Answer: B Explanation: When σ is unknown, we estimate it with s, giving the t statistic: (x/ - μ₀) / (s / √n). 44. The p-value is defined as: A) The probability that the null hypothesis is true B) The probability of committing a Type I error C) The probability of observing a test statistic as extreme as or more extreme than the observed value, assuming H₀ is true D) The probability of rejecting the null hypothesis Answer: C Explanation: The p-value measures the strength of evidence against H₀; a small p-value indicates strong evidence against the null hypothesis. 45. If the p-value is less than the significance level α, we: A) Fail to reject H₀ B) Reject H₀ C) Accept H₀ D) Increase α Answer: B Explanation: When the p-value < α, the result is statistically significant, and we reject the null hypothesis in favor of the alternative.
A) Reject H₀ B) Fail to reject H₀ C) Accept H₀ D) Conclude μ < 50 Answer: A Explanation: Since 0.03 < 0.05, we reject H₀ and conclude there is sufficient evidence to support μ > 50.
51. In a test of H₀: μ = 50 versus Hₐ: μ < 50, a sample yields a p-value of 0.08. At α = 0.05, we: A) Reject H₀ B) Fail to reject H₀ C) Conclude μ < 50 D) Conclude μ = 50 Answer: B Explanation: Since 0.08 > 0.05, we do not have sufficient evidence to reject H₀. We do not "accept" H₀; we fail to reject it. 52. A researcher tests H₀: μ = 20 vs Hₐ: μ ≠ 20 and calculates a p-value of 0.02. At α = 0.01, the conclusion is: A) Reject H₀ B) Fail to reject H₀ C) The test is significant D) Reject Hₐ Answer: B Explanation: Since 0.02 > 0.01, the result is not significant at the 1% level; we fail to reject H₀. 53. If a test statistic falls into the rejection region, we: A) Fail to reject H₀ B) Reject H₀ C) Cannot determine the outcome D) Increase the sample size Answer: B Explanation: The rejection region defines the set of test statistic values that are unlikely if H₀ is true; falling into it leads to rejection. 54. The critical value for a test is: A) The value of the sample statistic B) The boundary separating the rejection region from the non-rejection region C) The p-value D) The population parameter
Answer: B Explanation: The critical value is the threshold against which the test statistic is compared to decide whether to reject H₀.
55. For a one-sample t-test for the mean, the degrees of freedom are: A) n B) n - 1 C) n - 2 D) n + 1 Answer: B Explanation: The t-test for a single mean uses df = n - 1 to account for estimating one parameter (σ) from the sample. 56. Which assumption is NOT required for the one-sample t-test for the mean? A) Random sample B) Independent observations C) Normal population distribution (or large n) D) Known population variance Answer: D Explanation: The t-test does not require a known population variance; it estimates σ with s. The known variance case uses the z-test. 57. If the underlying population is heavily skewed and n = 15, what should a statistician do? A) Proceed with the t-test as usual B) Use a z-test instead C) Be cautious; consider nonparametric methods or collect a larger sample D) Always reject H₀ Answer: C Explanation: The t-test is robust to moderate skewness for larger samples (n ≥ 30). For small samples with heavy skewness, alternative methods are safer. 58. A test for a population proportion uses the distribution: A) t-distribution B) z-distribution (normal approximation) C) F-distribution D) Chi-square distribution Answer: B Explanation: The large-sample test for a proportion uses the z-statistic, relying on the normal approximation to the binomial distribution.
C) 99%
D) 99.9%
Answer: C Explanation: Confidence level = 1 - α, so 1 - 0.01 = 0.99 or 99%.
64. If the p-value is 0.045, the result is statistically significant at: A) α = 0.01 only B) α = 0.05 only C) α = 0.05 and α = 0. D) α = 0.01, 0.05, and 0. Answer: C Explanation: Since 0.045 < 0.05 and 0.045 < 0.10, it is significant at both levels. It is not significant at α=0.01 because 0.045 > 0.01. 65. Which of the following will decrease the width of a confidence interval for the mean? A) Increasing the confidence level B) Increasing the population standard deviation C) Increasing the sample size D) Decreasing the sample size Answer: C Explanation: Larger sample sizes reduce the standard error, resulting in a narrower interval. 66. Which of the following will increase the width of a confidence interval for the mean? A) Decreasing the confidence level B) Decreasing the sample size C) Decreasing the population standard deviation D) Increasing the sample size Answer: B Explanation: Smaller sample sizes increase the standard error, resulting in a wider interval. 67. A matched-pairs t-test is used when: A) Comparing two independent groups B) Comparing two means from the same subjects or matched subjects C) Comparing more than two groups D) Testing a single proportion Answer: B Explanation: In matched-pairs designs, each subject serves as its own control or subjects are paired, and the test is applied to the differences.
68. In a matched-pairs t-test, the null hypothesis typically states that the mean difference μ_d is: A) Greater than 0 B) Less than 0 C) Equal to 0 D) Not equal to 0 Answer: C Explanation: The standard null hypothesis is that there is no difference between the paired measurements: H₀: μ_d = 0. 69. If a 95% confidence interval for the mean difference in a matched-pairs design contains 0, we: A) Reject H₀ at α = 0. B) Fail to reject H₀ at α = 0. C) Conclude the mean difference is positive D) Conclude the mean difference is negative Answer: B Explanation: If 0 is plausible (inside the CI), we do not reject the null hypothesis of no difference. 70. The standard error of the mean difference for matched pairs is: A) s_d / √n B) σ_d / √n C) s_d / n D) √(s_d²/n) Answer: A Explanation: The SE for the mean of the differences is the sample standard deviation of differences divided by √n. 71. Which of the following is a valid alternative hypothesis for a two-tailed test? A) μ > 10 B) μ < 10 C) μ ≠ 10 D) μ = 10 Answer: C Explanation: The "not equal" sign (≠) defines a two-tailed alternative. 72. Which of the following is a valid alternative hypothesis for a right-tailed test? A) μ > 10 B) μ < 10
77. If the sample mean is 25, the standard error is 2, and the critical value z* = 1.96, the 95% confidence interval is: A) (23.04, 26.96) B) (21.08, 28.92) C) (23, 27) D) (25, 27) Answer: A Explanation: Margin of error = 1.96 × 2 = 3.92. Interval = 25 ± 3.92 = (21.08, 28.92). Wait, compute: 25 - 3.92 = 21.08, 25 + 3.92 = 28.92. Option B is correct. Let me re-check options: A is (23.04, 26.96) which would be 1.961=1.96, but SE=2, so E=3.92. B is correct. Let me fix the options conceptually in my head. Since I have to write Q77 properly: The correct interval is (21.08, 28.92). I will write that as the correct option. 77. If the sample mean is 25, the standard error is 2, and the critical value z = 1.96, the 95% confidence interval is: A) (23.04, 26.96) B) (21.08, 28.92) C) (23.00, 27.00) D) (25.00, 27.00) Answer: B Explanation: Margin of error = 1.96 × 2 = 3.92. Interval = 25 ± 3.92 = (21.08, 28.92). 78. If you want to reduce the margin of error to half its current value while keeping the same confidence level, you need to multiply the sample size by: A) 2 B) 4 C) √ D) 1/ Answer: B Explanation: Since E ∝ 1/√n, to halve E, we need √n to double, so n must increase by a factor of 4. 79. A researcher conducts a test and gets a p-value of 0.02. Which of the following is correct? A) The result is significant at α = 0.01 but not at α = 0. B) The result is significant at α = 0.05 but not at α = 0. C) The result is significant at both α = 0.01 and α = 0. D) The result is not significant at any α Answer: B Explanation: Since 0.02 < 0.05 but 0.02 > 0.01, it is significant at the 5% level but not at the 1% level.
80. When we fail to reject H₀, we conclude that: A) H₀ is proven true B) The sample evidence is insufficient to support Hₐ C) Hₐ is false D) The test has high power Answer: B Explanation: Failing to reject H₀ does not prove it true; it only means there is insufficient evidence to support the alternative. 81. Which of the following scenarios results in a larger t critical value for a given confidence level? A) n = 10 B) n = 30 C) n = 100 D) n = ∞ Answer: A Explanation: Smaller sample sizes mean smaller degrees of freedom, leading to heavier tails and larger t critical values. 82. The chi-square test for a single population variance is based on the assumption that the population is: A) Uniform B) Normal C) Binomial D) Skewed Answer: B Explanation: The chi-square test for variance is highly sensitive to normality; it requires the population to be normally distributed. 83. In testing the variance, the null hypothesis is often stated as: A) H₀: σ² = σ₀² B) H₀: σ² > σ₀² C) H₀: μ = μ₀ D) H₀: p = p₀ Answer: A Explanation: Tests for variance typically test whether the population variance equals a specified value. 84. The test statistic for a one-sample chi-square test for variance is: A) (n-1)s² / σ₀² B) (n)s² / σ₀² C) (n-1)σ² / s² D) (x/ - μ₀) / (s/√n)
