Docsity
Docsity

Prepare for your exams
Prepare for your exams

Study with the several resources on Docsity


Earn points to download
Earn points to download

Earn points by helping other students or get them with a premium plan


Guidelines and tips
Guidelines and tips

Example Multiple Choice Questions for Exam 4 - Statistical Methods | STT 201, Exams of Data Analysis & Statistical Methods

Material Type: Exam; Professor: Pathak; Class: Statistical Methods; Subject: Statistics and Probability; University: Michigan State University; Term: Spring 2009;

Typology: Exams

2010/2011

Uploaded on 10/23/2011

titans-gators
titans-gators 🇺🇸

5

(3)

37 documents

1 / 7

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
STT 201, Sections 7--11 UNIT 4
Example Multiple Choice Questions
1-4. For each of the following testing situations, give the p-value of the observed test statistic. Pick the closest.
1. Two-sample t-test of H0: 1 - 2 0 versus Ha: 1 - 2 < 0, sample sizes n1 = 12, n2 = 10, tobs = 1.38.
Adopt the conservative approach to the degrees of freedom.
(a) .10 (b) .05 (c) .20 (d) .50 (e) .90
2. Two-sample (large samples) z-test of H0: 1 - 2 0 v. Ha: 1 - 2 > 0 if p1 = .47, p2 = .33 and zobs =
1.15.
(a) .085 (b) .071 (c) .054 (d) .125 (e) .101
3. A categorical variable y has five levels. Two populations are tested for having the same distribution of y
using large samples drawn from each population and 2obs = 11.12.
(a) .085 (b) .071 (c) .054 (d) .025 (e) .037
4. A categorical variable y has three levels. Four populations are tested for having the same distribution of y
using large samples drawn from each population and 2obs = 2.20.
(a) .085 (b) .001 (c) .050 (d) .900 (e) .100
5-6. A researcher administered a reading comprehension test to each of 16 subjects before and after a special
course that each subject took to improve reading comprehension. The researcher set the hypotheses H0: d 0
and Ha: d > 0 where d denotes difference (improvement). The summary of the data is given below.
Descriptive Statistics
N Mean Standard deviation
Before 16 52.8 9.3
After 16 56.8 8.1
Difference 16 4.0 7.6
5. What is the value of tobs for testing H0?
(a) 2.11 (b) 7.6 (c) 0.45 (d) 2.00 (e) 4.0
6. What is the p-value of tobs?
(a) .10 (b) .05 (c) .025 (d) .01 (e) .005
7-12. A veterinarian believes that the rate of hip dysplasia in Boxers (breed of dog) is less than the rate hip
dysplasia in American Bulldogs (breed of dog) and decides to test the hypothesis H0: 1-2 0 v. Ha: 1-2 < 0. A
random sample of 80 Boxers has 8 dogs with hip dysplasia. A random sample of 140 American Bulldogs has 21
dogs with hip dysplasia.
7. What is the standard error of the estimate p1 – p2 used in calculating the CI for 1 - 2?
(a) .085 (b) .071 (c) .045 (d) .037 (e) .101
8. What is the 99% CI estimate of the difference in population rates 1 - 2?
(a) -.05 .12 (b) -.05 .07 (c) -.05 .09 (d) -.05 .20 (e) -.05 .17
9. What is the standard error of the estimate p1 – p2 used in calculating the z-test statistic? (Hint: This is the
denominator of z-test statistic.)
(a) .085 (b) .071 (c) .064 (d) .047 (e) .101
10. Compute the value zobs of the z-test statistic.
(a) 0.52 (b) -0.90 (c) 1.96 (d) –1.06 (e) -1.96
11. What is the p-value of the test statistic zobs that is appropriate for this test? Pick the closest.
(a) 0.30 (b) 0.18 (c) 0.14 (d) 0.05 (e) 0.025
12. What decision is reached at 2% level of significance?
(a) reject the null hypothesis (b) retain the null hypothesis
pf3
pf4
pf5

Partial preview of the text

Download Example Multiple Choice Questions for Exam 4 - Statistical Methods | STT 201 and more Exams Data Analysis & Statistical Methods in PDF only on Docsity!

STT 201, Sections 7--11 UNIT 4 Example Multiple Choice Questions 1-4. For each of the following testing situations, give the p-value of the observed test statistic. Pick the closest.

  1. Two-sample t-test of H 0 :  1 -  2  0 versus Ha:  1 -  2 < 0, sample sizes n 1 = 12, n 2 = 10, tobs = 1.38. Adopt the conservative approach to the degrees of freedom. (a) .10 (b) .05 (c) .20 (d) .50 (e).
  2. Two-sample (large samples) z-test of H 0 :  1 -  2  0 v. Ha:  1 -  2 > 0 if p 1 = .47, p 2 = .33 and zobs = 1.15. (a) .085 (b) .071 (c) .054 (d) .125 (e).
  3. A categorical variable y has five levels. Two populations are tested for having the same distribution of y using large samples drawn from each population and ^2 obs = 11.12. (a) .085 (b) .071 (c) .054 (d) .025 (e).
  4. A categorical variable y has three levels. Four populations are tested for having the same distribution of y using large samples drawn from each population and ^2 obs = 2.20. (a) .085 (b) .001 (c) .050 (d) .900 (e). 5-6. A researcher administered a reading comprehension test to each of 16 subjects before and after a special course that each subject took to improve reading comprehension. The researcher set the hypotheses H 0 : d  0 and Ha: d > 0 where d denotes difference (improvement). The summary of the data is given below. Descriptive Statistics N Mean Standard deviation Before 16 52.8 9. After 16 56.8 8. Difference 16 4.0 7.
  5. What is the value of tobs for testing H 0? (a) 2.11 (b) 7.6 (c) 0.45 (d) 2.00 (e) 4.
  6. What is the p-value of tobs? (a) .10 (b) .05 (c) .025 (d) .01 (e). 7-12. A veterinarian believes that the rate of hip dysplasia in Boxers (breed of dog) is less than the rate hip dysplasia in American Bulldogs (breed of dog) and decides to test the hypothesis H 0 :  1 -  2  0 v. Ha:  1 -  2 < 0. A random sample of 80 Boxers has 8 dogs with hip dysplasia. A random sample of 140 American Bulldogs has 21 dogs with hip dysplasia.
  7. What is the standard error of the estimate p 1 – p 2 used in calculating the CI for  1 -  2? (a) .085 (b) .071 (c) .045 (d) .037 (e).
  8. What is the 99% CI estimate of the difference in population rates  1 -  2? (a) -.05  .12 (b) -.05  .07 (c) -.05  .09 (d) -.05  .20 (e) -.05 .
  9. What is the standard error of the estimate p 1 – p 2 used in calculating the z-test statistic? ( Hint: This is the denominator of z-test statistic. ) (a) .085 (b) .071 (c) .064 (d) .047 (e).
  10. Compute the value zobs of the z-test statistic. (a) 0.52 (b) -0.90 (c) 1.96 (d) –1.06 (e) -1.
  11. What is the p-value of the test statistic zobs that is appropriate for this test? Pick the closest. (a) 0.30 (b) 0.18 (c) 0.14 (d) 0.05 (e) 0.
  12. What decision is reached at 2% level of significance? (a) reject the null hypothesis (b) retain the null hypothesis

13-15. A study is undertaken to compare the distributions of weights of adult perch in Lake Michigan, Lake Huron and Lake Erie. Random samples of sizes 100, 200, 300 respectively are obtained from these lakes and each sample fish is weighed. Weight is recorded in four categories. The sample data are shown below. Weight Category Under 8 ounces 8 – 12 ounces 12 – 16 ounces 16 ounces or more Total Michigan 10 40 40 10 100 Huron 45 65 60 30 200 Erie 65 135 80 20 300 Total 120 240 180 60 600

  1. In this application, the chi-square test is called (a) goodness-of-fit test (b) test of homogeneity (c) test of independence
  2. Which lake has the highest percentage of its sample fish in the 16 ounces or more category? (a) Michigan (b) Huron (c) Erie
  3. For computing ^2 obs in this application, the expected count for the cell (Huron, 8 - 12 ounces) and the contribution from this cell into ^2 obs are (a) 60, .42 (b) 65, 0 (c) 70, .36 (d) 75, 3.46 (e) 80, 2.
  4. Suppose that ^2 obs = 10.55. What is its p-value? (a) approx. 0.05 (b) approx. 0.001 (c) approx. 0.10 (d) approx. 0.02 (e) approx. 0. 17-22. Use the following MTB printout to test the significance of the slope of the regression line y ˆ^ =  0 +  1 x based on a sample of size 8. Regression Analysis The regression equation is y = 13.7 -1.47x. Predictor Coef SE Coef T P Constant 13.7444 5.291 ____ ____ x -1.466 1.012 ____ ____
  5. Give a point estimate of the slope of the population regression line. (a) 13.7 (b) –1.47 (c) 8 (d) 1.012 (e) 95%
  6. What is the value of the test statistic tobs for testing H^ 0 :^  1 ^0 v. H^ a :^  1 ^0? (a) -0.91 (b) -1.12 (c) -1.45 (d) 2.60 (e) 3.
  7. What is the p-value of the test statistic for testing H^ 0 :^  1 ^0 v. H^ a :^  1 ^0? Pick the closest. (a) .2 (b) .1 (c) .05 (d) .025 (e).
  8. For the test in #19, what decision is reached at 5% level of significance? (a) reject the null hypothesis (b) retain the null hypothesis
  9. At 5% level of significance, is this predictor important? (a) yes (b) no
  10. The 90% CI for the slope of the population regression line is (a) 1.012  1.966 (b) 1.012  2.848 (c) –1.466  1.9661 (d) –1.466  0.695 (e) –1.466  1.

Cost = 66.8 – 0.370age + 0.00583pages Predictor Coef SE Coef T p Constant 66.847 2.279 29.33 0. age -0.36963 0.05262 ______ _____ pages 0.005834 0.004334 ______ _____ s=4.279 R-sq = _____ R-sq(adj) = _______ Analysis of Variance SOURCE DF SS MS F p Regression 2 _____ 529.29 30.75 0. Error 50 860.71 17. Total 52 1919.

  1. Predict the price of the book that is 5 years old and has 500 pages. (a) 67.87 (b) 67.16 (c) 55.20 (d) 53.20 (e) 4.
  2. What is the sum of squares due to regression? (a) 55.2 (b) 1058.58 (c) 860.71 (d) 529.29 (e) 1919. 29.
  3. What is the percentage of variation in response variable explained by these two predictors? (a) 95% (b) 81.3% (c) 42.79% (d) 55.2% (e) none of these
  4. What decision is reached for the test of H 0 :  1 =  2 = 0 versus Ha: not so at 2% level of significance? (a) retain the null hypothesis (b) reject the null hypothesis
  5. Based on these data, what is the point estimate of the population regression coefficient  2? (a) 66.8 (b) -0.370 (c) 0.00583 (d) 2.279 (e) 0.
  6. The confidence factor for constructing the 95% confidence interval for the population regression coefficient  2 is (a) 2.009 (b) 1.96 (c) 2.021 (d) 2.000 (e) none of these
  7. The 95% confidence interval for the population regression coefficient  2 is (a) –0.36963 0.103 (b) 0.00583 0.014 (c) 0.00583 0. (d) 0.004334 0.103 (e) –7.02 0.
  8. The observed value of the test statistic for the test of H 0 :  2 = 0 v. Ha:  2  0 is (a) 29.33 (b) –7.02 (c) 1.35 (d) 30.75 (e) none of these
  9. The p-value for the test of H 0 :  2 = 0 v. Ha:  2  0 is (a) between .025 and .05 (b) between .05 and .1 (c) between .005 and. (d) between .1 and .2 (e) between .2 and.
  10. At 5% level of significance, is the predictor “pages” important? (a) yes (b) no METHODS BASED ON MATCHED SAMPLES: For matched pairs, treat the n differences as the sample. In this case, y^ is denoted by d , s is denoted by s^ d , and  is denoted by ^ d.

CI for ^ d : n s dt * d , where t *is based on n-1 degrees of freedom. Test statistic: n s d t d obs ^ , where t^ obs is based on n-1 degrees of freedom. METHODS BASED ON TWO INDEPENDENT SAMPLES Estimators: y 1 ^ y 2 for^  1 ^  2 , s 1 for^  1 , s 2 for^  2 ; p 1 - p 2 for^  1 ^  2 CI's (for  1 -  2 or  1 -  2 ) 2 2 2 1

  • 1 1 1 2

n p p n p p p p z

where isbasedon thesmallerof 1 1 and 2 1 (conservativeapproach to d.f.).

2 2 2 1 2

  • 1 1 ^2   t nnn s n s y y t Test Statistics ( for  1 -  2 or  1 -  2 ) where ,thepooledestimateofthecommonproportionunderH. 1 1 ( 1 ) 0 1 2 1 2 1 2 1 2 n n x x p n n p p p p zobs

where, with theconservativeapproach, isbasedonsmallerofn 1 - 1 andn 2 - 1 degreesof freedom. 2 2 2 1 2 1 1 2 obs t obs n s n s y y t

CHI-SQUARE TESTS:

1. Goodness of Fit Test Ho:  1 = 1o, …, k = ko v. Ha: Not so , where 1o, …, ko are the hypothesized

values. For large n, use the test statistic ^2 obs = (Oj - Ej)^2 / Ej where the summation is over all k cells and Ej=njo , j=1,…k. If Ho is true, then the test statistic has approximately a ^2 distribution with k-1 df.

2. Test of Independence Ho: The two classifications are independent v. Ha: Not so

One classification variable has r levels and the other has c levels. The observed frequencies Oj are in a contingency table with r rows and c columns. For large n, use the test statistic ^2 obs = (Oj - Ej)^2 / Ej where the expected cell frequencies are Ej = (row total)(column total)/(grand total) and the summation is over all rc cells of the contingency table. If Ho is true, then the test statistic has approximately a ^2 distribution with (r – 1)(c – 1) df.

3. Test of Homogeneity Ho: The populations are homogeneous with respect to the variable of interest

v. Ha: Not so. There are r populations and c levels for the classification variable. The observed frequencies Oj are in a contingency table with r rows and c columns. For large n, use the test statistic ^2 obs = (Oj - Ej)^2 / Ej where the expected cell frequencies are Ej = (row total)(column total)/(grand total) and the summation is over all rc cells of the contingency table. If Ho is true, then the test statistic has approximately a ^2 distribution with (r – 1) (c – 1) df. 2 2 2 1

  • 1 1 1 2

n p p n p p p p z

  1. c Margin of error=confidence factor* standard error=2.009*0.004334=0.0087, CI is 0.00583  0.0087.
  2. c t_obs=b2/SE(b2)=0.005384/0.004334=1.346.
  3. d Df=53-3=50, look up 1.35 on the t-table to find that the amount in right tail is between .05 and .1, p-value is between .1 and .2 (two-tailed test).
  4. b Since the null hypothesis was retained , there is not enough evidence to say that in the equation for the entire population predictor “pages” is multiplied by a coefficient that is not equal to zero. When multiplied by zero, predictor “pages” has no impact on response, and therefore it is not important.