( ) ( ) STAT 5031 Statistical Methods for Quality Improvement. Homework n = 8; x = 127 psi; σ = 2 psi (a) µ 0 = 125; α = 0.

Σχετικά έγγραφα
Μηχανική Μάθηση Hypothesis Testing

Statistics 104: Quantitative Methods for Economics Formula and Theorem Review

Math 6 SL Probability Distributions Practice Test Mark Scheme

Statistical Inference I Locally most powerful tests

Repeated measures Επαναληπτικές μετρήσεις

CHAPTER 25 SOLVING EQUATIONS BY ITERATIVE METHODS

Homework 3 Solutions

3.4 SUM AND DIFFERENCE FORMULAS. NOTE: cos(α+β) cos α + cos β cos(α-β) cos α -cos β

5.4 The Poisson Distribution.

HOMEWORK 4 = G. In order to plot the stress versus the stretch we define a normalized stretch:

Μαντζούνη, Πιπερίγκου, Χατζή. ΒΙΟΣΤΑΤΙΣΤΙΚΗ Εργαστήριο 5 ο

Lecture 34 Bootstrap confidence intervals

Potential Dividers. 46 minutes. 46 marks. Page 1 of 11

Aquinas College. Edexcel Mathematical formulae and statistics tables DO NOT WRITE ON THIS BOOKLET

Exercise 2: The form of the generalized likelihood ratio

Section 9.2 Polar Equations and Graphs

ΣΤΑΤΙΣΤΙΚΗ ΕΠΙΧΕΙΡΗΣΕΩΝ ΕΙΔΙΚΑ ΘΕΜΑΤΑ. Κεφάλαιο 13. Συμπεράσματα για τη σύγκριση δύο πληθυσμών

APPENDICES APPENDIX A. STATISTICAL TABLES AND CHARTS 651 APPENDIX B. BIBLIOGRAPHY 677 APPENDIX C. ANSWERS TO SELECTED EXERCISES 679

Μενύχτα, Πιπερίγκου, Σαββάτης. ΒΙΟΣΤΑΤΙΣΤΙΚΗ Εργαστήριο 5 ο

Other Test Constructions: Likelihood Ratio & Bayes Tests

Biostatistics for Health Sciences Review Sheet

Υπολογιστική Φυσική Στοιχειωδών Σωματιδίων

b. Use the parametrization from (a) to compute the area of S a as S a ds. Be sure to substitute for ds!

ST5224: Advanced Statistical Theory II

The Simply Typed Lambda Calculus

FORMULAS FOR STATISTICS 1

Instruction Execution Times

9.09. # 1. Area inside the oval limaçon r = cos θ. To graph, start with θ = 0 so r = 6. Compute dr

Statistics & Research methods. Athanasios Papaioannou University of Thessaly Dept. of PE & Sport Science

4.6 Autoregressive Moving Average Model ARMA(1,1)

Approximation of distance between locations on earth given by latitude and longitude

2 Composition. Invertible Mappings

Section 8.3 Trigonometric Equations

ΚΥΠΡΙΑΚΗ ΕΤΑΙΡΕΙΑ ΠΛΗΡΟΦΟΡΙΚΗΣ CYPRUS COMPUTER SOCIETY ΠΑΓΚΥΠΡΙΟΣ ΜΑΘΗΤΙΚΟΣ ΔΙΑΓΩΝΙΣΜΟΣ ΠΛΗΡΟΦΟΡΙΚΗΣ 19/5/2007

EE512: Error Control Coding

Επιστηµονική Επιµέλεια ρ. Γεώργιος Μενεξές. Εργαστήριο Γεωργίας. Viola adorata

Phys460.nb Solution for the t-dependent Schrodinger s equation How did we find the solution? (not required)

Solutions to Exercise Sheet 5

Econ 2110: Fall 2008 Suggested Solutions to Problem Set 8 questions or comments to Dan Fetter 1

Queensland University of Technology Transport Data Analysis and Modeling Methodologies

ANSWERSHEET (TOPIC = DIFFERENTIAL CALCULUS) COLLECTION #2. h 0 h h 0 h h 0 ( ) g k = g 0 + g 1 + g g 2009 =?

Inverse trigonometric functions & General Solution of Trigonometric Equations

Second Order Partial Differential Equations

Areas and Lengths in Polar Coordinates

Supplementary Appendix

Areas and Lengths in Polar Coordinates

2. THEORY OF EQUATIONS. PREVIOUS EAMCET Bits.

HISTOGRAMS AND PERCENTILES What is the 25 th percentile of a histogram? What is the 50 th percentile for the cigarette histogram?

6.3 Forecasting ARMA processes

the total number of electrons passing through the lamp.

Bayesian statistics. DS GA 1002 Probability and Statistics for Data Science.

derivation of the Laplacian from rectangular to spherical coordinates

χ 2 test ανεξαρτησίας

Section 7.6 Double and Half Angle Formulas

ΚΥΠΡΙΑΚΟΣ ΣΥΝΔΕΣΜΟΣ ΠΛΗΡΟΦΟΡΙΚΗΣ CYPRUS COMPUTER SOCIETY 21 ος ΠΑΓΚΥΠΡΙΟΣ ΜΑΘΗΤΙΚΟΣ ΔΙΑΓΩΝΙΣΜΟΣ ΠΛΗΡΟΦΟΡΙΚΗΣ Δεύτερος Γύρος - 30 Μαρτίου 2011

Problem Set 3: Solutions

Finite Field Problems: Solutions

Probability and Random Processes (Part II)

Right Rear Door. Let's now finish the door hinge saga with the right rear door

Theorem 8 Let φ be the most powerful size α test of H

1. Ηλεκτρικό μαύρο κουτί: Αισθητήρας μετατόπισης με βάση τη χωρητικότητα

Modbus basic setup notes for IO-Link AL1xxx Master Block

TABLES AND FORMULAS FOR MOORE Basic Practice of Statistics

η πιθανότητα επιτυχίας. Επομένως, η συνάρτηση πιθανοφάνειας είναι ίση με: ( ) 32 = p 18 1 p

Άσκηση 10, σελ Για τη μεταβλητή x (άτυπος όγκος) έχουμε: x censored_x 1 F 3 F 3 F 4 F 10 F 13 F 13 F 16 F 16 F 24 F 26 F 27 F 28 F

ο),,),--,ο< $ι ιι!η ι ηι ι ιι ιι t (t-test): ι ι η ι ι. $ι ι η ι ι ι 2 x s ι ι η η ιη ι η η SE x

Math221: HW# 1 solutions

2. Μηχανικό Μαύρο Κουτί: κύλινδρος με μια μπάλα μέσα σε αυτόν.

CHAPTER 48 APPLICATIONS OF MATRICES AND DETERMINANTS

ΕΙΣΑΓΩΓΗ ΣΤΗ ΣΤΑΤΙΣΤΙΚΗ ΑΝΑΛΥΣΗ

C.S. 430 Assignment 6, Sample Solutions

Lecture 2. Soundness and completeness of propositional logic

Για να ελέγξουµε αν η κατανοµή µιας µεταβλητής είναι συµβατή µε την κανονική εφαρµόζουµε το test Kolmogorov-Smirnov.

Αν οι προϋποθέσεις αυτές δεν ισχύουν, τότε ανατρέχουµε σε µη παραµετρικό τεστ.

An Introduction to Signal Detection and Estimation - Second Edition Chapter II: Selected Solutions

TMA4115 Matematikk 3

Second Order RLC Filters

Strain gauge and rosettes

Mock Exam 7. 1 Hong Kong Educational Publishing Company. Section A 1. Reference: HKDSE Math M Q2 (a) (1 + kx) n 1M + 1A = (1) =

PENGARUHKEPEMIMPINANINSTRUKSIONAL KEPALASEKOLAHDAN MOTIVASI BERPRESTASI GURU TERHADAP KINERJA MENGAJAR GURU SD NEGERI DI KOTA SUKABUMI

Ordinal Arithmetic: Addition, Multiplication, Exponentiation and Limit

1. Hasil Pengukuran Kadar TNF-α. DATA PENGAMATAN ABSORBANSI STANDAR TNF α PADA PANJANG GELOMBANG 450 nm

(1) Describe the process by which mercury atoms become excited in a fluorescent tube (3)

Εισαγωγή στην Ανάλυση Διακύμανσης

519.22(07.07) 78 : ( ) /.. ; c (07.07) , , 2008

τατιστική στην Εκπαίδευση II

w o = R 1 p. (1) R = p =. = 1

CHAPTER 12: PERIMETER, AREA, CIRCUMFERENCE, AND 12.1 INTRODUCTION TO GEOMETRIC 12.2 PERIMETER: SQUARES, RECTANGLES,

SCHOOL OF MATHEMATICAL SCIENCES G11LMA Linear Mathematics Examination Solutions

Δείγμα (μεγάλο) από οποιαδήποτε κατανομή

Pg The perimeter is P = 3x The area of a triangle is. where b is the base, h is the height. In our case b = x, then the area is

is like multiplying by the conversion factor of. Dividing by 2π gives you the

UDZ Swirl diffuser. Product facts. Quick-selection. Swirl diffuser UDZ. Product code example:

Homework 8 Model Solution Section

ECE 308 SIGNALS AND SYSTEMS FALL 2017 Answers to selected problems on prior years examinations

CE 530 Molecular Simulation

ΑΓΓΛΙΚΑ Ι. Ενότητα 7α: Impact of the Internet on Economic Education. Ζωή Κανταρίδου Τμήμα Εφαρμοσμένης Πληροφορικής

SECTION II: PROBABILITY MODELS

UNIVERSITY OF CALIFORNIA. EECS 150 Fall ) You are implementing an 4:1 Multiplexer that has the following specifications:

Transcript:

STAT 531 Statistical Methods for Quality Improvement Homework 3 4.8 n = 8; x = 17 psi; σ = psi (a) µ = 15; α =.5 Test H : µ = 15 vs. H 1 : µ > 15. Reject H if Z > Z α. x µ 17 15 Z = = =.88 σ n 8 Z α = Z.5 = 1.645 Reject H : µ = 15, and conclude that the mean tensile strength exceeds 15 psi. (b) P-value = 1 Φ(Z ) = 1 Φ(.88) = 1.99766 =.34 (c) In strength tests, we usually are interested in whether some minimum requirement is met, not simply that the mean does not equal the hypothesized value. A one-sided hypothesis test lets us do this. (d) x Zα σ n µ 17 1.645 8 µ 15.8 µ 4.1. (a) x ~ N(µ, σ), µ = 1, α =.1 n = 1, x = 1.15, s =.3 Test H : µ = 1 vs. H 1 : µ > 1. Reject H if t > t α. x µ 1.15 1 t = = = 1.5655 S n.33 1 t α, n 1 = t.1, 9 =.81 Do not reject H : µ = 1, and conclude that there is not enough evidence that the mean fill volume exceeds 1 oz. (b) α =.5 t α/, n 1 = t.5, 9 =.6 µ α n ( S ) µ ( S ) x t S n x + t S n α /, n 1 /, 1 1.15.6 1 1.15+.6 1 11.993 µ 1.37 (c) We can check this with a normal probability plot fillvol area score 11.96.5-1.64 11.98.15-1.4 11.99.5 -.67 1.1.35 -.39 1..45 -.13 1..55.13 1.3.65.39 1.4.75.67 fillvol Normal probability plot of fill volume 1.6 1.4 1. 1. 11.98 11.96-1.8-1.1 -.4.3 1. 1.7 score 1

1.5.85 1.4 1.5.95 1.64 The plotted points fall approximately along a straight line, so the assumption that fill volume is normally distributed is appropriate. For the confidence interval, the standard deviation is.3 oz, and the sample size is n=1, giving (n-1)s =.81. There are 9 degrees of freedom, and the upper and lower ½ % points of this chisquared are.7 and 19. respectively. So a 95% confidence interval is given by.81.81 ;.46 < σ < < σ <.3. The 95% CI for σ is.1 < σ <.55 19..7 4.15 x ~ N(µ, σ), n = 16, x = 1.59 V, s =.999 V (a) µ = 1, α =.5 Test H : µ = 1 vs. H 1 : µ 1. Reject H if t > t α/. x µ 1.59 1 t = = = 6.971 S n.999 16 t α/, n 1 = t.5, 15 =.131 Reject H : µ = 1, and conclude that the mean output voltage differs from 1V. (b) /, n 1 µ α /, n 1 µ x t S n x + t S n α 1.59.131.999 16 1.59+.131.999 16 9.77 µ 1.79 (c) σ = 1, α =.5 Test H : σ = 1 vs. H 1 : σ 1. Reject H if χ > χ α/, n-1 or χ < χ 1-α/, n-1. ( n 1) S (16 1).999 χ = = = 14.97 σ 1 χ α/, n 1 = χ.5,16 1 = 7.488 χ 1 α/, n 1 = χ.975,16 1 = 6.6 Do not reject H : σ = 1, and conclude that there is insufficient evidence that the variance differs from 1. (d) ( n 1) S ( n 1) S χ σ α /, n 1 χ1 α /, n 1 (16 1).999 (16 1).999 σ 7.488 6.6.545 σ.391.738 σ 1.546 Since the 95% confidence interval on σ contains the hypothesized value, σ = 1, the null hypothesis, H : σ = 1, cannot be rejected.

=.5; = = 7.69 (e) α χ1 α, n 1 χ.95,15 ( n 1) S σ χ 1 α, n 1 (16 1).999 σ 7.69 σ.6 σ 1.436 (f) The probability plot looks decently linear, so we are OK with the normality assumption. 4. n =, x = 18, ˆp = x/n = 18/ =.9 (a) p =.1, α =.5. Test H : p =.1 versus H 1 : p.1. Reject H if Z > Z α/. np = (.1) = Since (x = 18) < (np = ), use the normal approximation to the binomial for x < np. This gives Z ( x +.5) np (18 +.5) = = = np (1 p ) (1.1).3536. If we omit the continuity correction, we get x np 18 Z = = =.47 np (1 p ) (1.1) Z α/ = Z.5/ = Z.5 = 1.96 Either way, do not reject H, and conclude that the sample process fraction nonconforming does not differ from.1. P-value = [1 Φ Z ] = [1 Φ.3536 ] = [1.638] =.736 (b) α =.1, Z α/ = Z.1/ = Z.5 = 1.645 pˆ Z pˆ (1 pˆ ) n p ˆp + Z ˆp (1 ˆp ) n α / α /.9 1.645.9(1.9) p.9+ 1.645.9(1.9).57 p.13 4.1. n = 5, x = 65, ˆp = x/n = 65/5 =.13 (a) p =.8, α =.5. Test H : p =.8 versus H 1 : p.8. Reject H if Z > Z α/. np = 5(.8) = 4 Since (x = 65) > (np = 4), use the normal approximation to the binomial for x > np. 3

( x.5) np (65.5) 4 With the continuity correction this gives Z = = = 4.387, and np (1 p ) 4(1.8) x np 65 4 omitting the correction gives Z = = = 4.1 np (1 p ) 4(1.8) Z α/ = Z.5/ = Z.5 = 1.96 Either way, reject H, and conclude the sample process fraction nonconforming differs from.8. (b) P-value = [1 Φ Z ] = [1 Φ 4.387 ] = [1.99997] =.6 (c) α =.5, Z α = Z.5 = 1.645 p pˆ + Z pˆ (1 pˆ ) n α p.13+ 1.645.13(1.13) 5 p.155 4. (a) n 1 =, x 1 = 1, ˆp 1 = x 1 /n 1 = 1/ =.5 n = 3, x =, ˆp = x /n = /3 =.67 (b) Use α =.5. Test H : p 1 = p versus H 1 : p 1 p. Reject H if Z > Z α/ or Z < Z α/ x + x 1 + n + n + 3 1 pˆ = = =.6 1 Z pˆ pˆ.5.67 1 = = = pˆ (1 pˆ ) 1 n + 1 n.6(1.6) 1 + 1 3 1 Z α/ = Z.5/ = Z.5 = 1.96 Z α/ = 1.96.784 Do not reject H. Conclude there is no strong evidence to indicate a difference between the fraction nonconforming for the two processes. (c) ˆ ˆ pˆ (1 pˆ ) pˆ (1 pˆ ) 1 1 ( p p ) Z + ( p p ) 1 α / 1 n n 1 pˆ (1 pˆ ) pˆ (1 pˆ ) 1 1 ( pˆ pˆ ) + Z + 1 α / n n 1.5(1.5).67(1.67) (.5.67) 1.645 + ( p p ) 1 3.5(1.5).67(1.67) (.5.67) + 1.645 + 3 p p 1.5 ( ).18 4.5 Test H : µ d = versus H 1 : µ d. Reject H if t > t α/, n1 + n. 4

t α/, n1 + n = t.5, =.8188. d =.417; s d =.13 ( d ) t = d s n =.417.1311 1 = 1.1 ( t = 1.1) <.8188, so do not reject H. There is no strong evidence to indicate that the two calipers differ in their mean measurements. 4.8 The test is two-sided, so the cutoff value will be z.5 = 1.96. We want a probability of Type II error of.1 when the mean is actually. We can figure the needed sample size using the formula given in class ( zα / + zβ ) σ (1.96 + 1.8)3 n = = = 3.78. µ µ 15 4.3 (a) We wish to test H : λ = λ versus H 1 : λ λ. Select random sample of n observations x 1, x,, x n. Each x i ~ POI(λ). In other words, we need n=4 n x ~ POI( nλ ). i Using the normal approximation to the Poisson, if n is large, x = x/n = ~ N(λ, λ/n). Z = x λ λ n. Reject H : λ = λ if Z > Z α/ / (b) x ~ Poi(λ), n = 1, x = 11, x = x/n = 11/1 =.11 Test H : λ =.15 versus H 1 : λ.15, at α =.1. Reject H if Z > Z α/. Z α/ = Z.5 =.5758 Z = ( x λ) λ n = (.11.15).15 1 = 1.38 ( Z = 1.38) <.5758, so do not reject H. Lab portion 1. > x <- c(145, 15, 153, 148, 141, 15, 146, 154, 139, 148) > y <- c(15, 15, 147, 155, 14, 146, 158, 15, 151, 143) > t.test(x, y) Welch Two Sample t-test data: x and y t = -.778, df = 17.843, p-value =.459 alternative hypothesis: true difference in means is not equal to 95 percent confidence interval: -6.7943 3.1943 sample estimates: mean of x mean of y 147.6 149.4 The t test shows no evidence of a difference between the two groups i = 1. > n <- 16 > C <- sqrt(qchisq(.95, n-1) / (n-1)) 5

> C [1] 1.9886 So the needed C value is 1.9. Now we ll trace out the OC curve. I also draw a horizontal line at OC=. to help answer the followup question > sig <- seq(1,,.5) > OC <- pchisq((n-1)*(c/sig)^, n-1) > dev.new() > plot(oc~sig, main="oc curve for testing sigma", xlab="sigma", + ylab="oc", type="l") > abline(h=.) Rather than try reading the answer off the curve, I ll get R to tell me. > offset <- abs(oc -.) > sig[offset==min(offset)] [1] 1.55 The σ for which the OC is closest to. (ie the power is closes to 8%) is σ=1.55. As this is just a bit bigger than the 1.5 the question asks about, the needed sample size is going to be a bit above 16. Some searching by getting the right C and then evaluating the OC gives: > for (n in 16:) { + C <- sqrt(qchisq(.95, n-1) / (n-1)) + oc <- pchisq((n-1)*(c/1.5)^, n-1) + print(paste(n, oc)) + } [1] "16.55187981698" [1] "17.347835499885" [1] "18.1591476313886" [1] "19.19847491854759" [1] ".183648745149" So we needed n=19 3. The equation for the OC was given in class. Coding it up gives 6

mu <- seq(15, 18,.1) > sigma <- > n <- 8 > OC <- pnorm(1.645 - (mu-15)*sqrt(n)/sigma) > dev.new() > plot(mu, OC, main="oc curve for normal mean", xlab="mu", + ylab="oc", type="l") 4. I ve edited out some unnecessary lines > allin <- read.csv("http://stat.umn.edu/hawkins/531/table_4e11.csv") > > fitter <- with(allin, lm(satisfaction ~ Severity)) > summary(fitter) Coefficients: Estimate Std. Error t value Pr(> t ) (Intercept) 119.533 11.1673 1.73.9e-1 *** Severity -1.176.343-5.19 4.45e-5 *** Residual standard error: 14.95 on 3 degrees of freedom Multiple R-squared:.57, Adjusted R-squared:.5 F-statistic: 5.19 on 1 and 3 DF, p-value: 4.447e-5 The regression is highly significant. Each 1 unit increase in severity corresponds to 1.18 units less in patient satisfaction. The graph of the relationship looks fine; no evidence of the sort of violations that would get you worried. > dev.new() > with(allin, plot(severity, Satisfaction, main="table 4E.11")) > abline(fitter) 7

8