hw4 solution - 4 a The test statistics and p-Values for the...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 2
Background image of page 3
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 4 a The test statistics and p-Values for the three iraIiahles—added—in—order tests are: . ._. _ 1053.623 _ _ __fi 1 r (Jig—W — 30.762 (119.1,” P a 0.001. 17 .. .. 133.113 - =—= ' =:Z a: '3.l 1%ch Ills) 40128? 7.325 df. 1., 16 0.0]. P 0.0-) 16 iii Fig.3 Xg,Xs}= 33629382 =1.563 df: l, 15 [1.] =12 Pcil 0.25 - 15 X3. artng are significant; .1”; is not. b Ho.',fi'1=fi2 = 0 vs. HA: 3% e U and."orfi; = CI in the model Y=fl3 +3151} +3313 +fi513 + E 221.?25 Ffll’hX‘. X3) = = 4.577 dfzfi, 15 15 0.025 P-=: 0.05 At least one ofthe variables Xi and X2 is significant, given thatin is in the model. C Using the given ADIOVA table, we cannot test Hg: ,8; = ,3; = D in the model Y= ,8; + .3111] + ,3ng + ,83X3. + E, as the variables have not entered the model in the right order to perform this test. The test statistic that would be needed is: [Regression 55(Xl. X3 . X3) - Regression SSflijy ., FILE X. - i L' 3' -) ResidualMSthsz.X5} d X: and X; are important variables whereas X; should be excluded from the final model. . a The three F—tests are: i H0: ,6". =fi2 = 0 vs. H2: first! and-"orfi’zeO inthe model l"=,80 +£1.15 +39ng + E. F(.¥1,X2)=26.36 df: 2, 22 P-=: 0.0001 At a = 0.05, we reject Hg and conclude that ,81 .= 0 and-'orflg e 0. ii Hg: ,6'; = 0 vs.H_.:,:,81¢ 0 in the model Y=,3;+35’LX1 + E. FlCth=1J4 df: 1,23 P=0.2 At a = 0.05, we do not reject H0 and conclude thatfi‘;= 0. iii Hg: ,0" = 0 vs. H_.:,:fig e 0 in the model Y=fi:+x5’1X2 + E. Flngj = 53.96. df: l, 23 P-=: 0.0001 At a = 0.05, we reject Hg and conclude that ,8: .= 0. X3 is the more important predictor of J". In chapter 8, problem 3th), the model containing both predictors was chosen based on R] values. However, based on the partial F—tests above, we would choose the more parsimonious model containing only X2. 12 The two variables-added—in-order tests are: t See 03(0) ii H0: ,8: = U 0‘s. Haj: e U in the model Y=fl3+file+j93X3 + E. H E, I Y }_ 55601111 _ 85(X;Ji'.‘}—SSEX1} ‘ 3 ‘ 1 Residual 0130112,) Residual 0130212,) 102520.014? 402312202 = 4230 .2253 = “'45 22 df: . 22 P-=: 0.001 At a = 0.05, we reject Hg and conclude thatfig e 0 in the model 1"=.3:+.31X1+.3241'2+ 5- c The two variables-added—in-order tests are: i See 6a(iii}. ii Hg: ,8; = 0 0‘s. H_:.: ,8; .= 0 in the model }'=fig +IBIX1+fifill +E. a . _102570.814T—101932.665? _ _ M 2*(X1 Jig—W432: dil,“ 22 P:= 0.25 At 0: = 0.05, we do not reject Hg and conclude thatfi‘; = 0 in the model 3—: .3: +3154??st + E- d Source df 83 MS F X1 X2 1 638.149 638.149 0.328 X2 X1 1 92339.089 92339.089 42.45? Residual 22 42806225 1945.238 Total 24 145372.040 e No differences are discernahle; all approaches indicate that A] does not significantly aid in the prediction of 1’. either alone or in addition to X2. X3 does significantly aid in the prediction of I’. 8' 3 H0319? =33 = 0 “14:11:32 -= 0 Mid-"OWE = U in the model 1930 +135? +£2.15 +193X5+ E. {7010.03+l0.93) Page X.)=—3=32.79 df: 2, 21 2243.23 /21 P c=2 0.001 At ct = 0.05. we reject Ho and conclude that at least one .6,- == 0. b Ho:fiI=,33=0vs.HA: atleastonefli e 0 inthe model Y=fio+fl£fi+fisfi+£ [13953.04+]I‘010.03 . 7 = 2 = 2 -2 '1': PIER-"A; [l'[l.931+22-°l$.23}4j 1 6'4 dr— 1“ P c: 0.001 At at = 0.05. we reject hi] and conclude that at least one .63- = 0. c The F test corresponds to comparing the models 1’: fig — E and 11,330 +1911“. +1923} 433350 — E- 13. :1 Ho: .31 =13: = 0 vs. Hi: .6; .= 0 and-"0:131 e 0 in the model I’=.6t1 -,31X1-,3;.fi +E {XL = DWNCOSI. X; = URBAN) FLY]. X2} = 8.52 df: 2. 23 P=0.013" At ct = 0.05. we would reject Hg and conclude that at least onefij = 0. h 1 1:11:31 = 0 vs. Ha: .61 e 0 inthe model 1"=,6’: +19tXl-E. F _ Regression SSflXIJ _ 132.6203 _ [SSY - Regression SSflle].-" 24 _ [600.96l5— 132.6203] 324 df: 1.24 0.01 criPcri 0.025 At 0. = 0.05. we reject Hg and conclude that '31 .= 0. ii 210:3: = 0 vs. H32 .62 = 0 inthe model 1"='.-5’: +file— 6313+ E. FQLEl IQ = 8.21 df: 1.23 P=0.0088 At or = 0.05. we reject Hg and conclude that '33 e 0. =6.80 c i Ho: ,3: = 0 vs. Ha: ,8: .= 0 in the model l'=fi: +1395:— E. Regression SSflXl, X3) — Regression SS(X._ | X1) [SSY {Regression 8301'; . X: 3— Regression SSQ’ll X: )j] 24 = 255.??85—663318 =1105 [600.9615—{255.2235—66.33l3)]:"24 ' (if: 1. 24 0.001 {P c: 0.005 At a = 0.05, we reject Hg. and conclude thatfi'; .= 0. ii Hmfil = 0 vs. Hg: .81 .= 0 in the model }’=,6'g —fi1)1'1—fi}2'3+E. F011 '3) =4.42 df: 1, 23 P= 0.046? At a = 0.05. we rejectHc. and conclude that atfil e 0. d Source i SS MS F X1|X2 1 66.334 66.334 4.42 Xngg 1 123.153 123.158 8.21 Residual 23 345.133 15.008 Total 25 600. 962 9 Both predictors are neces sarjg. Each variable is significant regardless of the order entered in the model and both should be included. ...
View Full Document

This note was uploaded on 02/17/2012 for the course MPP 7510 taught by Professor Sounakchakraborty during the Spring '12 term at Missouri (Mizzou).

Page1 / 3

hw4 solution - 4 a The test statistics and p-Values for the...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online