- MOBILE PHONE FEATURE PREFERENCES AND CONSUMPTION PATTERN OF STUDENTS IN UNIVERSITY OF SARGODHA
- Twelve P Value Misconceptions
- IJMMR-V7N2-2014-6
- Week 10 Pt 3
- Business Research
- Regression Notes 01
- statistic
- DOE Engi 9516 Assignment
- Statistics Review 5 Comparison of Means
- skittles report
- Four Steps of Hypothesis
- chaid1.pdf
- Hypothesis Testing
- Lab 01- Scientific Method and Statistics (New Version)
- Hypothesis
- Twelve+P+Value+Misconceptions
- RegressionMemoDineshJayapathy
- Statistics
- IBNR reserves for health insurance actuaries
- Hypothesis Testing
- StatisticalTests (1)
- Ews
- Ch09 Estimation ConfidenceInterval
- Hypothesis Testing
- Week 6 - Hypothesis Testing
- BR_Lec 20 & 21
- Analysis
- New Proposal
- 1410.6966v1
- Estimatiom Ch 9
- Starters
- Nicknames banned from Aion
- Starters
- Muse - New Born
- Technicolor TG784nv3
- Solution-24

215

**Part III Review Solutions
**

ˆ III.1. (a) Use a two-sample z for proportions ( p1 is the proportion of the gingko group who believed they were taking gingko, etc.). (b) Use a two-sample t for means ( x1 is the mean score of the gingko group, etc.).

III.2. (a) Use a t test for means (we want to examine the mean of responses to the attractiveness question). (b) A matched-pairs test would be appropriate, because we should keep each couple’s responses together.

ˆ III.3. (a) Use a one-sample z for a proportion (with p as the sample proportion of workdays missed ˆ ˆ in Spain). (b) Use a two-sample z for proportions ( p1 and p2 are the proportions of missed workdays in Spain and Switzerland, respectively).

III.4. (a) Use a two-sample z for proportions. (b) Two-sample t for means. (c) Two-sample z for proportions. III.5. (a) Label the subjects from 01 to 44, and choose 22 subjects to eat regular chips first. From line 101 of Table B, we choose 19, 22, 39, 34, and 05. (b) Use a matched-pairs test for means (which is, of course, equivalent to a one-sample t test).

t* df Confidence interval III.6. (a) We were given x = 11.4 and s = 26.09 100 1.984 0.47581 to 8.92419 minutes, and for df = 100, we have t* = 1.984 134 1.97783 0.48890 to 8.91110 (software gives 1.9748 for df = 161).Therefore, 186.017 1.97280 0.49960 to 8.90040 the 95% confidence interval is either 11.4 ± 4.067 = 7.333 to 5.467 minutes, or 11.4 ± 4.048 = 7.352 to 15.448 minutes. (b) The standard error of the difference x f − xm is SE 2.12916. The table on the right shows three

possible choices of df and the intervals that arise from them. III.7. (a) The standard error of the difference x f − xm is SE 2.12916 (this was also found in the previous solution). The test statistic is t = (11.4 − 6.7) / SE 2.21, which is significant at the 5% level (see the table on the right for possible P-values). (b) t procedures are robust with large samples. III.8. SE =

( 0.07 )( 0.93)

2455 +

df 100 134 186.017

P-value 0.01478 0.01449 0.01425

( 0.14 )( 0.86 )

1191

0.01130; the 99% confidence interval is (0.07 – 0.14) ±

(2.576)(0.01130) = –9.91% to –4.09%. If we assume that there were 172 and 167 black child-care workers in each group, we can use the plus four method (which gives a similar interval): p1 0.0704 and p2 0.1408, SE 0.01132, and the interval is –0.0704 ± 0.0292 = –0.0996 to –0.0413. Because 0 is not in our 99% confidence interval, we would reject H0: p1 = p2 at the 1% level.

ˆ III.253 and t = 2.0123.0447 = 0. Using the plus four method: p1 = 0. p2 = 0. There is no evidence of a difference between the top melt and bottom melt biomasses. SE = 0. The appropriate critical value is 2. To test H0: p = 0.132 ( ) = 2.9 ± 0. .3352. p2 = 1077 0. and the resulting intervals. With the hypotheses H0: µ1 = µ2 vs. and the plus four interval is 0. z = 14.0241 = 0.0241 = 0. we compute p = 0. The P-value is 0.581 2. ˆ 775 III.10. for which P < 0. Using the plus four method: p = 1180 = 0.576 SE = 0. Because 0 is not in our interval. Ha: p ≠ 0. 839. (In fact.578 Confidence interval –0.2457 to 0. so we have no reason to doubt H0. We find that p = 904 1176 = 0. and the 95% confidence interval is 0. SE = 0.79 and P is practically 0. ˆ III.2905 ± 0. after rounding.42 /5 + 0. p1 = 840 680 ˆ 0.16. this gives a P-value close to 0.13.6314.7687.01735. We have very strong evidence that the proportion favoring mandatory registration was lower in 2001.2465 to 0.9− 2. we would reject H0: p1 = p2 at the 1% level.0578.4.3359.8 0.7678 ± 0. the TI-83 output on the right uses the more accurate df = 8. We have p1 = 0.50.8111. 0.7928.0123. This gives P > 0. we would reject H0: µ1 = µ2 at the 1% level.395 with df = 4. the 99% confidence interval is (0.00005.4 = 2. III.580 2.6314) ± 2.42 /5 = 0.12.7942 to –0. and the interval is 0.8526)/SE = –5. ˆ ˆ ˆ III.01737. Ha: µ1 ≠ µ2.9216 and p2 = 0. III. The standard error of x1 − x2 is SE = 2 2 s1 / 2455 + s2 /1191 0.70. and SE = ˆ ˆ ˆ ˆ p1 (1 − p2 ) p2 (1 − p2 ) + 840 1077 = 0.5 vs.4058 III.8526.9226 – 0. so the 90% confidence interval is 2.9 ± 2.5)/ = 1.4 5 Note: Watch out for students who might try to make something of the (coincidental) fact that.23. their associated t* values.7943 to –0.7678. (as in the previous exercise).5)(0.52).15. we have t = ( x1 − x2 ) / s1 / 840 + s2 /1077 = −0.7687.07533.132 (df = 4).8658.81 t* 2. For the hypotheses H0: µ1 = µ2 vs.7919. we compute SE = 0.) 2 2 III. which is not quite significant at the 5% level.7944 to –0. and gives P = 0.11. Because 0 is not in our 99% confidence interval.14. Then SE = 0.6311.5) 250 z = (0.253 = − 0.4056 –0.90. or 1777. Ha: µ1 ≠ µ2.01606 and z = (0.4057 –0.216 Part III Review df 1000 1190 2456.3 ng C per liter.7446 to 0.7687 – 0. The table on the right shows three possible choices of df.7687 ± 906 0.56 − 0.7437 to 0. SE = 0. Regardless of the df used (100. and p = 904+1024 1176+1201 = 0.9226.9.5. the margin of error is the same as the standard deviation.56 and (0.5 to 3.

19.56 ± 0. and the plus four interval is 0. we can apply the plus four method: p1 = 0. Because the sample size is so large. with slight differences because 0.0085 = 0.6400 ± 0.52 and p2 = 0.39.6485.7397 ± 0.Solution 217 140 250 ˆ III.64 ± 1.01755. we test H0: p = 0.1652 to –0.4980 to 0.00435.0551 = 0.0282. . SE = ˆ ˆ ˆ ˆ p1 (1 − p1 ) p2 (1 − p2 ) + 863 12. SE = 0. The TI-83 output on the right gives rough confirmation of this. with slight differences because 0.0314. we have SE = 0. The TI-83 output on the right gives rough confirmation of this. and the t statistic is t = (5. SE = 0.6201. The test statistic is z = (0.56.5.175. ˆ ˆ (b) With p1 = 0.0096.96 SE = 0.7950.03782. SE = 0.5591 ± 0.5202 and p2 = 0.078716. p = is 0.64 were entered as 449/863 and 7794/12178. With p = 0.52 − 0. and the 95% confidence interval is 0. so the 95% confidence interval is 0.0611 = 0.21. we have strong evidence that the mean remating time is longer for large spermatophores.52 and 0. Ha: p1 < p2. p = 0.6314 to 0.0615 = 0.5) 863 = 1. III. Ha: µ1 > µ2.15 – 4.52 was entered as 449/863.8283)/SE = –2.0312. Using the plus four method. Regardless of whether df = 19 or 32.64.4.0281. so P > 0. III. We cannot conclude that more than half of Americans hold this opinion. SE = 0. p = = 0.1649 to –0.17. ˆ III.7832.7358 ± 0. this is strong evidence that VLBW graduation rates are lower.0746.7358.7909. we ˆ compute p = 179+193 242+233 = 0.6215.6400.0553 = 0. SE = 0. the P-value is tiny.33)/SE = 10.7397 and p2 = 0. and the plus four 246 interval is 0. ˆ III.22.8283.64 was entered as 7794/12178.20. To test H0: µ1 = µ2 vs. (b) The high ˆ ˆ school graduation rates are p1 = 0.01755) = –0. (a) The sample sizes (especially for Europeans) are so large that even a small difference between Americans and Europeans would be significant. we find that the standard error of the difference x1 − x2 is SE = 0.6844 to 0.6315 to 0. and the 95% confidence interval is 0.0452 = –0.64. The one- sided P-value is P = 0.0748.6400. = and the plus four interval III.6485.576)(0.5)(0.5591.52. Using the plus four method.1198 ± 0. 178 = 0. with slight differences because 0.00435. The TI-83 output on the right gives rough confirmation of this.7397. and the interval is –0. (a) This is an observational study (one cannot “assign” a baby’s birthweight). If we use those counts. so the 99% confidence interval is (0.7397 – 0. Ha: p > 0. p = 181 = 0.64) ± (2. and SE = 0. SE = 0. = 0.01753.18.6807 to 0.34. the plus four method gives 7796 almost exactly the same interval: p = 12182 = 0. To test H0: p1 = p2 vs. (c) ˆ With p = 0. and z = (0.5)/ (0.10.5 vs.4985 to 142 254 0.52 – 0.

105. we find SE = 298.20 using either df = 82 or df = 128. and compute p1 = 54 = 0.28.15. or 216. For comparing IQs.39) ± t*(0. so we have strong evidence that walking and resting flies respond differently.1153 and t = –0.1253.31 to 0. so P > 0.32 To test H0: µ1 = µ2 vs.35. (b) No: SE = 65. and the t statistic is t = (87.06057.02786. multiply by n.79. (a) Not very significant: SE = 0.1250)/SE = 6. we find that the standard error of the difference x1 − x2 is SE = 2.” stands for standard error (of the mean). III.6042.39 1.39 ± t*(0. and z = (0. With either df = 29 or df = 56. and p = 126+124 = 0. all choices of df give t* close to 1.1). III.04—significant at α = 0. The summary statistics are n 30 30 Nitrites 9 5 22 8655 32 97666 322110 87555 44 8 2 Control 4 5 5 6 6 7 7 8 8 9 9 10 5 4 56 01124 6888 011123 5678 00013 9 3 x 7807. There are no extreme outliers. though the nitrite group may be slightly left-skewed.2 – 89.11) = 0. t = –1.08. Regardless of whether we take df = 100.1250.29. The test statistic is t = (86.207 to 0.01 g/day (regardless of the value of t* used.4.4194.59397 0. There is n Drivers Conductors 98 83 Calories s x 2821 2844 435.6636 (df = 82).25. III.7)/SE = –3.8438 and p2 = ˆ p= 54+4 64+32 4 32 = 0. (c) (0.0005. III. .8438 – 0.58 437.8.23. and the test statistic is z = (0. Ha: µ1 < µ2. e. No: You have information about all California counties (not just a sample). Both stemplots are reasonably symmetrical.4194)/SE = –2.573 grams—whether we use t* = 1.218 Part III Review III.2937. To test H0: µ1 = µ2 vs.50.24 0. or 247.26. ˆ 64 ˆ III.2937 – 0. Ha: µ1 ≠ µ2.24. (a) “s. p2 = 0.30 8072. This has a two-sided P-value of P = 0. 123.50 using either df = 82 or df = 173. there is little evidence that nitrites decrease amino acid uptake.3560.9. we test H0: (1 = ) p2 vs. III. so P > 0. The P-value is tiny.73370. To find s. and compute SE = 1. Ha: p1 ≠ p2. = 0.05.1253) = –0.16 and t = –0.664 (df = 80) or t* = 1.93 Nitrites Control s 1235. Then SE = 0. For the proportion using drugs.30 Alcohol s x 0. the P-value is less than 0.27.0188—a significant difference.00215 little evidence that calorie consumption is different. For any choice of df (100.4.89. (b) 0.20 1068. P > 0.08. Ha: µ1 < µ2.8)/SE = –2. We test H0: (1 = ) p2 vs. Ha: p1 ≠ p2. we test H0: µ1 = µ2 vs.24 – 0.10588. we have strong evidence that VLBW men have lower IQs. and compute 37+52 ˆ ˆ ˆ p1 = 0. SE = 0. this t has a two-sided P-value of about 0.6 – 94.20.292).

22 = 0. so t = 0.3345 to 5.016 (df = 43. (c) The observed differences between the two groups of mice were so large that they would be unlikely to occur by chance if the two groups were not different. Using t* = 2. The insulin means are x1 = 5. and s = 5. Ha: µ1 > µ2.59. with margin of error t*s/ 29. With df = 85.48 to 17. this gives –14. .3.632 ng/ml. with a slightly low outlier of 4. and we took many samples. 5. if the groups were the same. the 95% confidence interval is x ± t * s / n = 14. so 0. especially among pets.19 mg/dl and t = 1. Some possible answers are given below. and that underlying distributions are normal.0698 95% 5.060)(68/ 26 ) = 165.85 and s2 = 0. the difference in glucose levels would be so large less than 5% of the time. Doing so is overkill in this case because we do not know the correct sample sizes anyway.53 to 220.0 % C. the evidence would be even stronger.4479 serves as our best estimate of the earth’s density.922 = 5.57 to 52. but you do need them if you want to find the more accurate estimate for df. Minitab output Confidence Intervals Variable Date N 86 Mean 15. 17. For the data in Table I. SEM = s / n (b) Two-sample t-tests were done.9 − 0.3781 to 5. and the difference in insulin levels would be so large less than 0.32. It is unlikely that we have random samples from either population. There is nothing to keep us from using the t procedure.15 using either df = 22 or df = 43.2 0. The estimated standard error of the difference is SE = 0.) Note that you do not actually need to compute s1 and s2 to find SE.3).001 (for a two-sided alternative)—an even stronger result than the paper claimed.1134 III. This is not enough evidence to reject H0. and the standard deviations are approximately s1 = 0. we cannot conclude that pet cholesterol levels are higher.9 10 = 2.3. Table C allows us to say that P < 0.31.744.34.0840 99% 5.30.2 10 = 0. (And with larger sample sizes. 48 49 50 51 52 53 54 55 56 57 58 8 7 0 6799 04469 2467 03578 12358 59 5 Confidence Level Confidence Interval Margin of error 90% 5.33.47 mg/dl. the interval is 193 ± (2. the interval is – 13.909 SE Mean 0.5613 0.57 mg/dl.64 to 51. III.2209. using t* = 2. (a) “SEM” stands for “standard error of the mean”.477. (c) With df = 25. (b) The interval is (193 – 174) ± t*SE.75 ng/ml. we find SE = 16.922.64 mg/dl.744 StDev 5.074 (df = 22).17. x = 15. because wild-type and aP 2− / − mice are separate (independent) groups. (d) We are assuming that we have two SRSs from each population.3639 to 5.637 ( 95. Specifically.5177 0. or May 3 through May 6.01 days from April 20. With either df = 9 or df = 9.011) III.10 < P < 0. we find n = 86. The stemplot shows the distribution to be fairly symmetric. We have stronger evidence of a difference in insulin levels.I. III.92 + 0.9 and x2 = 0.909 days.9.88. x = 5. 14.5320 0. (a) To test H0: µ1 = µ2 vs.5% of the time.4479 and s = 0.Solution 219 III.

so we have strong evidence of a difference.07 g and the standard deviation is s = 19.05) that the standard deviations differ.41 g. The test statistic is F = 26. this means x ≥ 37. this means we reject H0 when x ≥ t * (108 / 50) = 15.19 (table) or c = 49.07 ± 2.01.26 g.(c) The power against µ = 100 is P( x ≥ 37) = P( Z ≥ 37 −100 ) = P( Z ≥ −4.05.70.82 to 61. where we find that 2.46.31 g.660 (using df = 100 from the table) or 1. Ha: σ1 ≠ σ2. it has a very small Pvalue. the second is acceptable.65 g. (c) The approximate power against µB – µA = 100 is 50 − 100 P( xB − x A ≥ 50) = P Z ≥ 4002 + 4002 350 350 This test will not often detect a difference of $100.092 = 15. III.69 ± 2.69 and s* = 14. Ha: σ1 > σ2. and discarding the outlier may cause us to underestimate the actual mean. and large sample sizes do not help. with a high outlier of 123 g. = P( Z ≥ −1. we get x * = 51.423 (conservatively) or 2.388 with df 25 and 22.38. III.403 (approximately) from Table C. so this should be quite adequate.045s / 30 = 46. We test H0: σ1 = σ2. 2 3 4 5 6 7 8 9 10 11 12 3 4899 12222455689 017789 089 12 26 III.025 < P < 0. a 95% confidence interval for the mean city particulate level is 54.2 with df 161 and 6.220 2 Part III Review III. 350 350 this means xB − x A must be above either c = 50.39 We have 30 available observations.043). x ≥ 36. (a) We test H0: σ1 = σ2.86 (software). If we discard the outlier. The first and third of these are the best choices. we take t* = 2.649 (df = 349 from software).048s * / 29 = 46. This F value can be compared to Table D values for df 120 and 100.73.12 to 57.65) = 0. Confidence intervals should be used with caution here: The outlier makes t procedures suspect.12 < F < 2. This F value can be compared to Table D values for df 20 and 20. and the interval is 51.36. (b) Reject H0 if t ≥ t*.9998. but for a power computation. III.405 (using software). The mean is x = 54.12) > 0. (a) With df = 49. For the three values of t* given above. or 2. A sample of size 50 will almost 108 / 50 certainly detect a difference of $100. This is significant evidence (at α = 0.05 (software gives 0.35. so xB − x A > t * 4002 + 4002 . the distribution is rightskewed. so 0. Because t = x /( s / n ). (a) We use either t* = 1. (b) Reject H0 if t ≥ t*. and x ≥ 36.37. (b) No: F procedures depend on Normality.69 134. respectively. 3 .2735t *. The test statistic is F = 682 442 = 2.

963 + 0.060 s / 26 = 0. The point (108. t = 2.1%. If we naïvely 0 01111111 use the t procedures in spite of the outliers.000 g. The distribution of differences (city minus rural) has 26 –0 11110 observations. and t = 2. even though we have removed from the data the two strongest individual pieces of evidence against H0. 1 1 If we throw out those outliers. (This is the solid line in ˆ the plot. we predict y = 93. there is only one observation — (51. and 0. Using the other 24 observations. s = 0 2222222 4. Ha: µc > µr is 0.) . the data seem to be more suitable 1 5 for a t procedure. 69) — which deviates from the pattern slightly.Solution 221 –0 32 III. 123) is a potentially influential observation (although it does not seem to deviate from the pattern of the other points).000 ± 2.087 g.297 to 4.111 to 1.1% of the variation in the data. we get x = 2. The plot (right) shows a strong positive linear relationship. We 1 8 can still reject H0. ˆ The regression line y = –2.) When x = 88 g. and y = 89.691 g.106 g. Computing the regression line ˆ without this point gives y = 1.889 g. The P-value for testing H0: µc = µr 0 5 vs.65 g. There are two high outliers (15 and 18).02 — pretty strong evidence against 0 7 H0.192 g.0935x explains r2 = 95. (This is the dashed line in the plot.41.02.38 (df = 25). III. we find x * = 1 1. A 95% confidence interval for µc – µr is 1. the scales should be the same on both axes.580 + 1.01 < P < 0.33 (df = 23). s* = 2.01 < P < 0.9942x and r2 = ˆ 92.069s* 24= 0. A 95% confidence interval for µc – µr is 2.40. Because both variables are measured in grams.192 ± 0 2.45 g when x = 88 g.

- MOBILE PHONE FEATURE PREFERENCES AND CONSUMPTION PATTERN OF STUDENTS IN UNIVERSITY OF SARGODHAUploaded byasifch_lums
- Twelve P Value MisconceptionsUploaded byderridaJacques
- IJMMR-V7N2-2014-6Uploaded byAngeliqueAquino
- Week 10 Pt 3Uploaded byJane Doe
- Business ResearchUploaded byAtif Iqbal
- Regression Notes 01Uploaded byDarfElfDrizzt
- statisticUploaded byNurul Jannah
- DOE Engi 9516 AssignmentUploaded bynuvan
- Statistics Review 5 Comparison of MeansUploaded byManish Chandra Prabhakar
- skittles reportUploaded byapi-271901299
- Four Steps of HypothesisUploaded byP3 Powers
- chaid1.pdfUploaded bypriya
- Hypothesis TestingUploaded byKalyan Dubal
- Lab 01- Scientific Method and Statistics (New Version)Uploaded by13ucci
- HypothesisUploaded bysiddhant
- Twelve+P+Value+MisconceptionsUploaded bybijugeorge1
- RegressionMemoDineshJayapathyUploaded byacephalist
- StatisticsUploaded byNurul Ekawati P
- IBNR reserves for health insurance actuariesUploaded bymoxb
- Hypothesis TestingUploaded byrns116
- StatisticalTests (1)Uploaded byAugustine Elinwa
- EwsUploaded byusman96
- Ch09 Estimation ConfidenceIntervalUploaded byYusuf Sahin
- Hypothesis TestingUploaded byImran Ahmad Sajid
- Week 6 - Hypothesis TestingUploaded byJeff Green
- BR_Lec 20 & 21Uploaded byshahidul0
- AnalysisUploaded byMillecai Macaranas
- New ProposalUploaded bySneha
- 1410.6966v1Uploaded bysaeed
- Estimatiom Ch 9Uploaded byDamian Deo