Download Final Exam Form A with Answer Key - Statistical Methods | STAT 303 and more Exams Data Analysis & Statistical Methods in PDF only on Docsity!
STAT303: Secs 507-
Spring 1998
Final Exam
Form A
Instructor: Julie Hagen Carroll
Name:
- Donât even open this until you are told to do so.
- Be sure to mark your section number (507, 508 or 509) and your test form (A, B, C or D) on the scantron!
- Sign your name where indicated on your scantron and PRINT your name at the top of this exam whether you sign the bottom or not. If you sign at the bottom, you are agreeing to have both this exam grade and your grade in the course posted by the last four digits of your Social Security Number.
- There are 25 multiple-choice questions on this exam, each worth 5 points. There is partial credit. Please mark your answers clearly on the scantron. Multiple marks will be counted wrong.
- You will have 2 hours to finish this exam.
- If you are caught cheating or helping someone to cheat on this exam, you both will receive a grade of zero on the exam. You must work alone.
- This exam is worth 120 points, and will constitute 30% of your final grade.
- Good luck!
Final grades will be available by 8AM FRIDAY. The Statistics office WONâT have them, however, so I would like my test score and final grade posted by the last four digits of my Social Security Number. This means it will appear on the Statistics Lab bulletin board no later than Friday at 8AM, and I wonât call the office asking about it since I know they have caller ID and Iâll lose a whole letter grade by calling!!! (Do you feel lucky?)
Signature:
SSN:
- What is the best conclusion based on the graph above?
A. Fail to reject H 0 at the 10% level and con- clude that the means are equal. B. Fail to reject H 0 at the 10% level and con- clude that the variances are equal. C. Fail to reject H 0 at the 10% level and con- clude that the means are not equal. D. Fail to reject H 0 at the 10% level and con- clude that the variances are not equal. E. Reject H 0 at the 10% level and conclude that the variances are not equal.
- Suppose we want to find out in general whether seniors, as a class, have a higher GPR than fresh- men, as a class (NOT whether they improve, like the last test). Which of the following would be the best method to test this hypothesis?
A. Take two independent random samples, one of freshmen and one of seniors, and compare their average GPRâs. B. Take two independent random samples, one of freshmen and one of seniors, and calcu- late the proportion that had higher GPRâs. C. Take a random sample of seniors, find the difference in their GPRâs as freshmen and as seniors, and compare that to zero. D. Take a random sample of seniors, find their average GPR and compare that to zero. E. Take a random sample of seniors, find their average GPR and compare that to their av- erage GPR as freshmen.
- Why did we decide to use α = 0.10 for Barlettâs test for equal variances?
A. It doesnât matter what α level you use as long as itâs greater than the F p-value. B. a Type I error is more critical since it means that weâre using an invalid F test. C. a Type II error is more critical since it means that weâre using an invalid F test. D. We want to reject as often as possible since that shows the treatment is significant. E. We want to reject as often as possible since that why we do hypotheses tests.
- Suppose the grades on this exam are normally distributed. If I give you z scores rather than points out of 100, what does it mean to get a z score of 1.5?
A. It means that you got 1.5 more points than the average. B. It means that you got 1.5 more questions right than the average. C. It means that you only missed 1.5 questions. D. It means that you only got 1.5 questions correct. E. It means that you are in the top 10% of the class for this exam.
- Aggies have been winning a lot of Big 12 Cham- pionships lately, so letâs see how they do in an- other contest. We want to determine which of the Big 12 schools is the least snooty. Weâre going to say snooty means that you drink high dollar beer rather than domestic stuff. (Lone Star, Pearl and Shiner get extra credit.) Since we know not all students drink, weâre going to compare the proportion of snooty beer drinkers out of beer drinkers in general at each school. We have to use proportions since none of the schools have the same number of beer drinkers. What type test should we use assuming we can say the proportions will be approximately normally dis- tributed?
A. 12 one-sample z tests, since weâre assuming normality B. 6 two-sample z tests, since weâre assuming normality C. ANOVA F test, since there are multiple groups/schools D. Chi-squared (Ï^2 ) test, since there are mul- tiple groups/schools E. independent t tests since we donât know the standard deviations
Source | Part SS df MS F Prob > F -------+------------------------------------ A | 125.778 2 62.89 11.32 0. B | .055556 1 .0556 0.01 0. A*B | .444444 2 .2222 0.04 0. Error | 66.6667 12 5. -------+------------------------------------ Total | 192.944 17 11.
- What conclusions may be made from the output above?
A. Since there is not significant interaction, we cannot make conclusions about either main effect. B. At the 10% level, the interaction is not sig- nificant, but factor A is. C. At the 1% level, the interaction is not sig- nificant, but factor B (just barely) is. D. Both the interaction and factor B are sig- nificant at the 5% level. E. Exactly two of the above are valid conclu- sions.
Number of obs = 22 F( 1, 20) = 3. Prob > F = 0. R-squared = 0. Adj R-squared = 0. Root MSE = 5.
Source | SS df MS ---------+------------------------------ Model | 98.7282903 1 98. Residual | 616.77171 20 30. ---------+------------------------------ Total | 715.50 21 34.
y | Coef. Std. Err. t P>|t| ------+----------------------------------- x | .6797128 .3798848 1.789 0. _cons | 21.86972 2.349109 9.310 0.
- What is the correct conclusion based on the out- put above?
A. At the 10% level, we can conclude that the xâs are useful in predicting the yâs. B. At the 1% level, we can conclude that the xâs are useful in predicting the yâs. C. At the 10% level, we can conclude that the true population slope is 0. D. Exactly two of the above are correct con- clusions. E. All of the above (excluding D.) are correct conclusions.
- Using the information above, what is the resid- ual, e, for the point (3,25) assuming this point is within the range of the data? A. -5. B. 1. C. 23. D. 19. E. 5.
- Let X ⌠N (12. 4569 , 2 .3425). Which of the fol- lowing probabilities is the largest? (Note: you donât need to do any calculations!) A. Pr(X = 18.2831) B. Pr(X < 11 .4127) C. Pr(X > 15 .2114) D. Pr(X = 4.5829) E. Pr(X > 10 .5918)
- What is the approximate correlation between the x and y displayed above?
A. strongly negative B. moderately negative C. weak D. moderately positive E. strongly positive
- If we add the point (5,30), which of the following will happen? A. The correlation, slope and intercept will be the same as before. B. The correlation and slope will change, but intercept will stay the same. C. The correlation will be the same, but the slope and intercept will change. D. The correlation and intercept will change, but the slope will be the same. E. The slope will change, but the intercept and correlation will be the same.
- Which of the following indicate the data is skewed to the right (positively)?
A. the mean is greater than the median B. the normal quantile plot curves down (both tails are below the line) C. the boxplot has a small box on top (above) a long line (whisker) D. All of the above indicate skewedness to the right. E. Exactly two of the above indicate skewedness to the right.
- Which of the following is/are true about the data represented above?
A. Conclude at the 1, 5 and 10% levels that the mean difference is 0. B. Conclude at the 1, 5 and 10% levels that the mean difference is not 0. C. Conclude at the 1, 5 and 10% levels that the mean difference is less than 0. D. Conclude at the 1, 5 and 10% levels that the mean difference is greater than 0. E. Conclude at the 10% level only that the mean difference is less than 0.
- What assumptions are necessary for the test above to be appropriate?
A. The data must have come from a normally distributed population. B. The variances are unknown, but equal. C. The data must have been randomly chosen. D. All of the above are necessary E. Exactly two of the above (excluding D.).
- Let X ⌠N (4, 22 ). What is xâ^ such that P (X > xâ) = 0.70?
A. 0. B. -0. C. 0. D. 5. E. 2.
- What is the best method of determining if the data is normally distributed?
A. Bartlettâs test B. F test C. histogram D. z test E. a normal quantile plot
- Why do we do Two-Way ANOVA rather than One-Way ANOVA twice?
A. You can save âmoneyâ by using the same data, whereas two One-Way ANOVAâs would need two sets of data. B. You can test more than one factor, i.e., two factors, simultaneously. C. You can test the interaction between two factors. D. All of the above. E. None of the above.
