Regression Analysis ~::::======:::::::::::::::=:::::::::::::=::::::::;~~~====~=== tNiRODUCTION 5,1 We know that the correlation studies the relaf h' b . hies X and y In this h II consid th I tons tp etween two vana · pter, we s a k er e re ated problem of prediction or estimation of the value of one ch~iable from a no~n v~l~e of other variable to which it is related. When there are_ two viiaria . bles X and• y and tf y is. influenced by X, i·e1 ., ·r y depends on X, then we get a simple hnear 1 v"ression or s1mp _e regres.s1on equation of 'Y on X'. Here y is known as dependent variable or re;ression or _exp lamed va~table and X is known as independent variable or predictor or explan~tor. ~~~case of a sunp le regression model if y depends on X, then the regression line of y on X is given b( Y = a + bX ... (A) st }-{ere a an~ b arc two con ~nts and their values are obtained by solving normal equatio.ns. th J-Jere a is e mtercept an~ b is the slope of the line. They are also known as regr~sSion 9111eters. Furthermore, b ts also known as the regression coefficient of Y on X and is also ::~oted by bP. A_nother regression line in which X depends y is : X = a + bY. Here b is the ession coeffic1ent of 'X on y , and is denoted by b ~r ~ . 1011 Regress !.'hows a relationship between the average values of two variables. Thus ,essiull is very helpful in estimating and predicting the average value of one variable for a '~~ell be made with the help of a gt1 mlue. of the I . other I I variable. Tire estimate or predictio11 mau J regression /me w uc,, s wws tire average value of one variable x for a given value of the other ,,ariable y. The best ave~age value of one variable associated with the given value of the other t'ariable may also be estm,ated or predicted by means of an equation and the equation is known as Ree.oression equatio11. s.2 TYPES OF REGRESSIONS Simple Regression. The regression analysis confined to the study of only two variables at a time 1s called the simple regression. Multiple Regression. The regression analysis for studying more than two variables at a time is known as multiple regression. Linear Regression. If the regression curve is a straight line, then there is a linear regression between the two variables under study In other words, in linear regression the relationship between the two \'ariables X and Y is linear. In order to estimate the best average values of the two variables, two regression equations are required and they are used separately. 011e equatio11 is used for estimating the value of X variable for a given value of Y variable and the second equation is used for estimating the 1•alue of Y variable for a given value of X variable. In both cases, the assumption is that one is an independent variable a11d the other is a dependent variable and 1•ice 1·ersa. 5.3 LINES OF REGRESSIONS A line of regression is the line which gives the best estimate of one variable X for any given \'a)ue of the other variable Y. 5.1 fcSJS {i, bat aot y) and if p aadb = l!..xbP .. q sip, Le., either botb essioa coefflden are neeative. IILll[ii=-=~11a l!!I" ,; J., jP 1111 lutH the su,e ..... •zl./tfal/lwlfto• tJ,e rep ,.._ - - . . 111¥ ,,.,,.,&ldiu to e,td, =.Ji.ixoM = .Jo.368 =0.606. a. Here :t -2+3+2+ .. 4+9+4+ re,t:~ M 1.7 IIITtt00 o, L,!AST SQUARES ,tell! Jr,/a /I /J used for obtalr,ir,g tJie ~ ~ TIN MeeW el Lalt iJ a ,nat/rlflllJ of a cww wl,ic/s fiJJ bat tlJ a gtW'I set of oblerY"''°;;bt 1qa•rts of dlrferencet bet ' It IS bated OD the . . .mptJoa tb•t tbt ,11111 0' the d•t• if ,ninfmum. The 0 atlrMted v•ltNt •ad the •daal oblerved value; to tbt ~ea data and consist, le \ method h wed to ol,taJa the belt ntda& 1traf&bt :.ent variable for dependent va,~'tbe bat flttJac ,tralpt line to the values of lodt9' bt obtained by means or I Mathematic.ally, the bttt ntdn& ,tral&bt 1111;1:nWbtll the relatlon1hlp l~ht~'1 - . ie-.."'"- ~ Ille.-" •bow• "hldl the rdadon t,etwttll the two vari\ainin, the 1,ett valutt or the con,,1 • ~ llnnr, the method of least .quaret Is uttd for ob f the constants in the rnunatrng llta i_ appropriate equation. for ol,tauung the beSI val°:n~ in the esiunating equation, arc 1 ,et of equauons, depending on the number of cons ed for obtaiffiffK the values of cocq...,,~ I11 be 10lved Tise syim" of equat/JJffS required be JO " 'h1, -~ItoUnuatlOIII• t/fe ntu,,llling equ111/JJ11 l• knowff as,N or.. - ""f t'on for fitting a rtv "' 1 The method of least ..quares 11 used in obtaining !he equa ~ ~~ti,,.'- "''k,,i ~- 5.7.1 Direct Method E.qua11on of a wa1ght line 1s : J' • 11 + bx ts a and b arc: Nor,,,al Equations for obtairung the values of the con,uin {/) l: y • Na + b!. x and aix + b!.x1. x • Independent variable, where y • Dependent variable, I• Sign ,,f summati""• a • Con,tant, (11) 1:xy • b - Conitsnt,Of . Of Ob N • Number pairs ·r he equation of the beil fit fs the ,tralght line : Y • a + bx, where the v:.luc:1 l>f a and h are obtained by ,,,lving the norm.al equatlom. •ervaurJn, • The method Ii illustrated by the following er.ample. t..rample 5. Pit a Jtra1ghl fine of Yon X from the following data --X L-_r.:..:_ O I 2 3 4 5 -;,--.._. 3 1 _:2 _ _ _:_1 _ __::.3:___ _ _ 2_ _ _4_ _ _ __ _ _ -.::."?--- --........ Solutfon. The regrtisi"n line of Yon X: Y•a+bX. The two normal equations to evaluate a and b are • I Y• aN+hIX I:XY aT,X + bI:)(l. y X 2 I 3 2 4 3 5 () I 2 3 4 5 6 .EX• 21, I: Y =20, xi 0 CJ J v.... -11 ... , (n'. 1 rhe equation of regression line Oxonyls: f x-x = b (y "' or X - - Y) CJx - -x = r-(yCJ y;;-\,. ,Ex:arnple 6. Calculate the re-e,s ,,. · ton 'coefficients fi =r . I: X 50, I: Y = 30, I: XY = I 000, I: X ! the fol~ng informatum: 5 1ut10n. Regression coefficient or y . . 3000, on X 15 given by : t Y • = /80• N. JO• 0 Table : Computation of Fitting a Straight Line r h~"' ---.!.::_ ro t JC)'-11i; 5irnilarlY, XY () I I 4 9 16 25 36 I: X2 =91 6 b,x= ,CJ'=I){Y-NXY CJx r,xz-N(X)z 6 Here 16 15 30 IXY X = tx _50 . . N -io=5, Subst1tutmg these Values m (I), we get ; 74. - l:Y 30 Y = N=10 =3. b = 1000-10(5){3) 1000-150 yx 3000-10(5)2 = 3000-250 -0.309. ... (I) iS ~ ~ T e s t QuanMative s;s ~ . tiJ)g (iii) in (,J, we get: 3x + 2 (3l 3x - l2x = 26 -6:r)•~ So"5ptil . o coeflideot or X on Y JS, given by • R,cresao -y- 1000-10(5)(3) 1000 - 150 -0497 IXY-YX _.:...:.----:--=--.. bx,= IY1-N(Y)z - 1800-l0(3)2 1800-90 . . ll d _,1 record the following data are ava,lable : Eumple 7. In a parM y estroyeu sr 1111'5· ~ . ( ) Sol Ution. 0 1,1" • x and y lllcd ~ • 7 = 22 .•. OJ and 64 x-45y = 24. Multipl}iog the equation (1) by 45, we get : 225x-45y = 990. Putting x = 6 in ( l ), we get : Hence x = 6 and y = 8. 30- y .•• (2) ... (lJ = 966 ⇒ x =6. Subrractmg (2) from (3), we get: 161.i tJJe p-0iot of mtenection of{').... = 22 => y = 8. p 2 . A gam . . regression equation of x on y ⇒ . lS : 8 64 y = --+-x. 15 45 Jte'l'"red correlation coefficient is given by: ~( ' · t~b~z, = = , .. b ii th le 9. for I 00 students of a class the • ~ aep e, IO' wOI also buepdvel is:aJJ]p . regresm,n equatio ., ks ,n commerce ( Y) is 3Y- 5X + 180 = o. Th . n o, maria in statistics (X) on # ,nark5 in statistics is (4/9Jth of the variance of em;::. maria ,n commerce iJ 50 and varuuu:e of,,,a~ and the coefficient of correlation between rna L ~ commerce. Find the mean maria in siansr1cs marll.J m the two subjects. solution, Given 3y - 5x + 180 = O or 3y + 180 = 5r L,et •x' represents ~rlcs m Statistics and 'y' represents marks in Conmerce. W}!en y = 50, x will be given by 5x = 3 (50) + 180 22 1 Sx - y = 22 or x = 5 + Y. 5 Calculation of Coefficient of Correlation: b b x, x, == - s· ·. bxy = 0.6 ; .!. . 15 ::::) r But we know that : b TX = is' = r CJ Y (J => 64 45 8 CJ y =iiX5 ::::) (Jy 40 =)= 13.JJ Hence. the standard de,iation of y : l3.3J. E~ample 8. Given rhe regression lines as 3x + 2y = 26 and 6x + y = 31. x andFind y. their point of intersection and interpret 11. · Also find the correlation coefficient between Solution. 3x + 2y = 26 ... (z) 6.r+y=31 ... (it) From (ii), y = 31 _ 6x ... (iir) 330 :66, 5 ax a, r~ = r o, ~ (given). Thus 0.6 ~9 = r [~) ~9 ⇒ 0.6 = r ( ½) ⇒ 2r = 1.8 or r = + 0.9. = 5• 64 byx =45 ' 8 Also (JX = , -½· Regression coefficient of x on y from the given equation is: 5X = Jy + 180 ⇒ X = 0.6y + 36 (+ve sign with r is taken as both the regression coefficients bx, and by,: are positive.) (c) ~ow it is given that vanance of x = o 2X = 25 X= Hence, the mean marks in Statistics are 66. 8 Hence, the coefficient of correlation r = ±vl-¾Jl-¼) 5x = 150 + 180 or :;::, 64 b)'X = -45 • r = ±Jbx, xbyx = ±)~ x~ = 15 . But . ,, .... .,u., 6 (b) The regression equation of y on x is : 64 24 64x- 45y = 24 or y = - x - 45 45 • (U) II ..,_ ltJ (4, "' lie on the regression lines and are obta· soh·ing the given regressmn equaoons. 5i-y k tbe fbUS• The mean nlues -62-9.t. regression line of y on x be • h ' +Z1•1'Le., l•U - -x 3 2 at of x ODY be : 6x +y = 31 ; e .. 31 l aodt.. . ., .. = -6- - , b 3 1 6 b :::: _ - and bx,= - - 1tl Variance of x = 25. · .r n y· 5x - y = 22. Regression equa11on o, x o · · .r on ... 64x - 45y = 24. Regression equa11on o, Y .,.. . Find (a) mean i·alues of x and y; (b) coefficient of correlation between x and y. (c) Standard deviation ofy. . .. ~. tiJ)gx:::::4m(11,J,weget: 1 • 31 ~Le~~•4;. .1,dJ(l.l • -24•7. 7 Example 10. The equations of the regression lines between two variables are exprened as lX - 3Y = 0 and 4 Y - 5X - 8 = 0. Find X and Y. the regression coefficients and the correlalion coefficient between X and Y. Solution. The two regression equations are : 2X - 3Y = 0. -CJ> and 4 Y - 5X - 8 = 0. - (M) The pomt ( X, Y) hes at the intersection of both the lines of regression equabGIII, ~ lwo regression equations. (1) x 4 ⇒ 8X-' 12Y = 0 (i1) x 3 ⇒ - l 5X + I 2 Y - 24 = 0 common proflclenc sis Test: Quantitative assume that the line •et Us ~ ... 1Oy "' Bx + 66 ==> l,\' 3>' . (). Also - ~r - o :2-4 ) ... , ' ,AlS'' f (---7 -... 40x = 18y + 214 ,knowthatr = l)llf \\IC n' (yos If tl:t• \\iriulll'<' <!/) ,:,- / 5, Jind tht• stand,ml dt'1·iat1011 ofx. Solution. Stllnng the given equations, we get x = 3 and .I' = I. • ,. ~ ~,,,,., , · 1 -- =- ftf fC sign is taken because both b 2 'i • 7 7 .-• ,Jl',,x ' -5 = - -, 7 IO 2 3y+ 3 : therefore, bx,,= - 6. [ ·: bf). 1s negative, so r 1s negative) =- {w =- /5. {ii fii Also ox+ 2y = 2y = 20 => .r = - '\o" r=- - ~. xb09 J -I (Hi)-J I/¥' . CJ ~ 'is b = r-'-=>-=lcr, ~ = ✓7 = 2,646 Agam xcr.r - => crx = - - x vl5 ' v D, w CJ." 3 21 CJr 3 5 Hence the standard de,i ation of x is 2.646. Example 12. Given that the i:iriance of x = 9 and the regression equations are 8t- /Oy + 66 = O. 40.i:- 18y = 214 Find (a} mean \'Glues ofx and y; (b) coefficient ofcorrelation berween x and _i, (c) standard del'iatio11 of y. Solution. (a) The mean values of x and y, i.e., (x, y) he on both the lines 8x-lOy = - 66. •.. (I) 40x- I8y = 214. Multiplying (I) by 5 and subtracting it from (2), we get ; 32y =544 or y =17. Substituting y = 17 in (I), we get: x = 13. Hence, the mean values of x and y are 5 20 Jr/ -9 •• -6 (i 100=+-. 10 and 6~ are ,-..uye.) -.1..a... x = 13, y = 11. •.. (2) (Given) ~wx-=>a,. =4. _ owing data: X = 20, y = I5, cr = ., · · tain the regression equations and find th·' " • a>• - 3• r = 0•7· . Ob . e most like/11 val 0.r Y. I, solution, The regression coefficients are : :J ue , • w en X = 24. £%11 a,. .r .. 1. ,~ )',t l cr>' 4 6 b"' = r-=>-a,. AlsO . cr.r 5 3 e the standard deviation ofy Is 4 J-fcllC , mple J3. Two variables gave thefioll . 3_ 2.1 - = o.s2s and b = CJ, 4 4 IC1 r -=0.7x-=0933 :,: 3 ' ' 0x The regression equation of X on y: x _ X = b (Y -Y) ::::> X - 20 = 0.933 (Y _ 15) ,,, 933 or _ ~ = o. y - 0.933 x 15 + 20 or X = 0.933Y + 6. JJence the regression equation X on y is : x = 0_933 y + , 6 b,.r. v = .::. _.:... \, Therefore, b1.• 48 ==>t-' cr.t = 9 ::::> cr.t = 3• 1 .S S ~, r .. o71 (r is posltiw since b1.1 und b . are po . ' • · •'ii Sttj , ..• r : • hf l ' /I ,, • _:,.. J ':' :, "' -t.5 'e) '-' I 11 F' I tit . ,11l' lll rnl11t•1· of two random mriablt's x and y and the Correl ~.:ump,, . mi < • • • . ' .• , b ,. a11 <<'<1fkicnt h<'fWf<'JI them wht·n the• nm /inl'.1· of n-gt't.'sS1011 ,m: gn t" J. o~ .fa + iy _ 2: = O,md 6x + 2y - 20 ~ 0. •·'I IV 21 x~-.H-! J;;;;--;;;; =!R;9 -><-:: .$ 5 Hence .\' • J 11nd ==> ny Is 18 40 7 • •20 11 I }"';" • - 2,28(1, Tims .\: • - -'•" I 10 + , .. - , t t,c regression equation or x O l'•- ~ --- 2.ltl(I. Q 'f ::= r - (1 = 07 . X4 The regression equation of Yon X: y _ ~ y = byx (X - X). Y - 15 = 0.525 (X - 20) or Y = 0.525X + 15 - 20 When X = 24, the most likely value of y is : x 0.525 or Y =0.525 X + 4.5, f = 0.525 X 24 + 4.5 = 17,1, Example 14. From the data given below estimate the most likely height of a brother whose sisters height is 70 cm Brother's · mean height is 67 cm with a s.d. of 3.5 cm Sister's. mean he1glu is 65 cm with a s.d. of 2.5 cm The coefficient of correlation between the heights of brothers and sisters is + 0.8. Solution. Let x, y be the variables corresponding to the heights of sisters and brothers. :. Then x = 65, ji = 67, cr, = 2.5, crr = 3.5, r = 0.8. rcr,. (0.8) (3.5) I.l2. b,.x= - - 25 CJX , The equation of line of regression of y on x ls: Now y-y = byx(x - x) ⇒ y- 67 = 1.12(x-65) ⇒ y=l.12r-U ,. ll)i ,,\,,\ \' '''lllP\,.,r-,, l \ t 1.1 l '" (l) '' l•>\tt)IN "'(I) (I•> IN '" l1J, 's l \>~)) 1()1 ►~ ,, ' t I IJ /(I (1/> \1 ' \•,.1: ,.,,o.~I' , ''" /'/ll'i l 1) 1111c/ l"\llllll'k I>, r , c,/>t,11111,/ l"' <I/'/ o;,,.,,, , , •11,1 /t' 111111') 11 / h l ' I I ,. , ' \' \' , , 1' I '() ,, • • I' • ' (/ ' \ /(' th,· ,,,,,,. ,,w1d ,. n/111111 " ' • • , " hr 11· c, 1111 I 1 , •1 111111 I' \ ' (I) , I I(, / ~01111 ''"· 1II' l\'l'll'SSl\111 " 1 \) I I 1()/, 1 I'() j/ I C, 1·1111nl11 11 " .. u " , d /, 1.0 I 11111 I ( ') \\'\' fl'( ' ,, " ·" 1 11 S1,lvi11~ (ll ' • ' II II I 1.01 -, , . ' j II I' I' 1111 \ ,~ : .J' ' II 1'1<:111111111 111 Il• 17• , • (:'/\'I'll 1/tc• /P//Pll 'll/1; I Sc'l'/1'1 ,,,.,.,,,~,· S1111ul,ml I 'nwu,,,, Soh1ll1111, I lr11• f 11, "' ,,., ,. /II /f I, I 'II 11h, ·It /' ,. ,.,, ' n,,,.,.,,.,_,.,, " · /,J(, t \' II ,,, 11'1 p/ / \ /(, 11111/1 I ••in· 11 liv 111 11 I I' 111 1• ,•· , I \( )11 I .... ss1, lh1, ( 'i ~ m•111~1I) II 1(,() I(,\ u I' 11·/, 1•11 I /) vu 1 • I 1,, [I I I(,, \' II I I'// I~ I ); l•l,K, 1.s. () 1/1) .\' 11,HH ""' 1,111 1(1\ 12 "·' 144 1-44 1()0 17 289 I , jl) 15X I t,7 ~ 64 !'i l'/0 ~ 25 170 56 25 I) .. (I ltiO I~() " I •I 170 (; l~ ,,,, 17~ 172 5 0 3 1/1 m 0 I 11 }.; 1/1 • l(t•. I' Ill () m ,. r ., 1/ll I I~ / ,/ 8 I (, ,/\: 1(, / I 7, (, '" ... ;~ 17(} 11, ~ •~. " · 1/11 1'11/11 1• I ',l//1/11//1 I' \ \ \(/(l, ''. (1 • . ' •I /\',\:'''"' • 1 Pf "·' 1/w /pl/p11·111,1: ""'" 11~•1r "'1'1' 11 { l'l, ll }; ,,,, .. ~2(, t I'• 16110 225 25 25 0 9 0 ~ ttly• 60 4'$ 0 JO 0 al ollcloncv Tosi: Quon . n_ vo - P c~.J-_;J..-_ 9,12 ------- j( . 10 • r.dwli• (l:1fr!_J~1/y)/N hll · - -r.,1l p;~d IN . u1:~1• h,.1 ~ 168,8; Y IX X lkre N 10, ( 12)x( 60)__/lo_ ~ 09 - ----- - ~ - - O 806 ( 60) 11 , 409 72 i2)2 526 (- -(t1M(r.,1yl1-_!t'_ - t,/x1 (f.1L\·)2/N ° · v }:)' N ~11,1~ 1.13 1690 s • ~(X.:..x )z -10--169, XY 409 72 337 806 360 - 446 "'0,756, 337 s2614.4 no 'P\!!~Ue St 337 siT6 .,,,c,e " r 0 ,659, ~~ - ,, n h (Y-Y) 'J' Rcgrrsslon ~:quntlon f X on Y ' ,\ - ,\ ) X 168 8 0.756 (Y 169 or ) 75 6)' r 168 127.764 ⇒ X= 0.756 }' +4 or ,\' 0 756Y + 168 (0. 756 ) (169 0· _ °'236 r. y ji ,.. b _(X X) Regression •:quation of yon , . J'.1 y I69 0.659 (X 168.8) y o. 65 9x + 169 - (0.659) ( 168.8) or Y 0.659X + . 57 76 :henX "' 164, y 0.659 x 164 + 57.761 108.076 + 57.761 or Y= 165.837. I I 510 VARIABILITY IN REGRESSION ANALYSIS . Let X and y denote the estimated (or computed) val~es. of X and y obtained from lhc c . c Then we know that the sum of the deviations of Y's and X's from th elr regression equations. regression lines are zero. Therefore l: (Y- Ye)'= Oand l: (X - Xe)= O Variability: Variability is the measure of spread or scatter of actual values around th, computed values Ye. It 1s given by (Y-Yc)2 Variability - - - 11 . . . _ (X-Xc) 2 or Variab1hty - - - - II STANDARD ERROR OF ESTIMATE Tire standard error of estimate is similar to sta11dard deviation. The sta11dard error of estimate is a measure of the variatio11 or scatteredness about the li11e of regression, whereai sta11dard deviatio11 measures the variatio11 or scattered11ess about the arithmetic mean. Standard error of Estimate indicates how precise the prediction of Y is based on Xor conversely. There are two types of standard error of estimates. (a} Standard Error of Estimate Yon X. It is denoted by Srx· It ts denved from the formula 5.11 1:(Y-Yc}2 2 1:Y -al:Y-bIXY or S = . , - - - - - - YX n-2 rx (11-2) where a and b are to be obtained from the normal equations and a = intercept, b = slope of line. S Also = Srx= cry ✓t-r 2 , where r is the coefficient of correlation X and Y; Srx predicts the values of Y based on X. (b) Standard Error of Estimate of X on Y. It 1s denoted by S .. It predicts the values ofX based on Y n ~ 11-2 sXY .. 0 XVr--:11-. rl2 O = (_ s • Jtx 2 -otx-btxY or b w b< rhcevaluatcd Iron, then lip d or111a1 equatJ " ) • " 2 ons ao d 11 • Intercept, b • slope of line. ,,AIS tiIe coefficient of correlation betw ' r 1s een X and Y. ~ . e)(PLAINED A.ND U~EXPLAINEo VARIATIONS 6'1Z c.,pta jpcd Variation. It ts the variation wh·tc h 1s. exp! · db v". giVCP by 1: (Y.C - f)2 • aine Ythe variable X. JI JS • • • lained Variations. It 1s the variati·o h. V11e%P n w tch is u I· . • _ due to some other factors (variables) tli . nexp atned by the variable X. This 0n 1s · a ecttng the t ta! · • . v,riau is . b !: ( y _ y )2. o vanatton m Y-values. c _ 11 given Y 2 Total Variation: 1:(Y -Y) = r [(Y -Y. ) (Y. - 2 J,JoW C + c-Y)] == !: (Y-Yc)2 +t (Yc-Y)2 +2L(Y-Y. )(Y.,-Y) == 1:(Y-Yc)2+1:(Yc-Yl c c == Un-Explained Variation + Explai· d " . . ne .-ar1atton. . "' . Thus, Explamed .-anance = 1:(Yc-Y) 2 n COEFFICIENT OF DETERMINATION ; Unexplained Variance= 1: (Y -Ye) n 2 513 · TIie coefficie~,t of determi1tatio~ is th e sq11are of the coefficient of correlation. It is equal to 1 • t'ton 111 • y • l · db~ r or r,J· The maximum . value of r2 ts unity and in that case all th e varta 1s exp ame the variation in X It is also defined as : . . Explained Variance Coefficient of Determmatton = --:::-----Total Variance · . D . . Unexplained Variance Coefficient of Non- eternunation = ::------Total Variance It IS denoted by k2• Also k1 = 1 - r Coefficient of Allienation. It is the square root of coefficient of non-determination. It is denoted by k. 5.14 UTILITY OF REGRESSION ANALYSIS 1. TIie cause and effect relations are indicated from the study of regression analysis. 2. It establishes the rate of change in one variable in terms of the changes in another variable. 3. It is usefttl i11 economic a11alysis as regression equation can determine an increase in the cost of living index for a particular increase in general price level. 4. It helps in prediction and thus it can estimate the values of unknown quantities. 5. It helps i11 determining the coefficient of correlation as: r = ✓b,x x bxy • 6. It enable us to study the nature of relationship between the variables. 7. It ca11 be usefttl to all natural, social and physical sciences, where the data are in fu11ctional relationship. Anal : :.:.si~s_ _ _ _ __ ·c,·ency Test: Quantitative Aptitude Common Pro f1 5.14 ,e~ botb (A) and (B). (C) None of these. MULTIPLE CHOICE QUESTIONS · Write the correct answer out ofthe given o,,es·. (A) measuring the extent of association between two variables. (B) establishing a mathematical relationship between two vanables. (C) predicting the value of . the dependent variable for a . given value of the independent vanable. (D) Both (A) and (C). 2. If there are two variables x and Y, then the number of regression equations could be: (A) 2 (B) I (C) Any number (D) 3. 3. The method applied for deriving the regression equations is known as : (A) C"ncurrent deviation (B) Least squares (C) Product moment (D) Normal equation. 4. What are the limits of the two regression coefficients ? (A) Must be positive. (B) No limit. (C) One positive and the other negative. (D) Both positive or both negative. 5. Two regression lines coincide when : ~)r=O ~)r=2 (C) r =+I 6. The two lines of identical when : (A) r = 1 C r=O (D) None of these. regression become (B) r = - I (D) (A) or (B). - . <lit' difference between the observed ~= ~= the estimated value in JS known as : (BJ error :!:'f:!$SJD.!: ~-sis s. In the line Y = 19 - (5/2) X, b i ~(A) 5/2 J:1 s equ ~ (B) 15/2 {C) - 5/2 . (D) None of the 9. In the equat10n X = 35/8 - (2/5) y se. ,b equa I to : l}IS (A) - 2/5 (B) 2/5 (C) 7/12 (D) 5/2. to. The line X = 31/6 + Y/6 is the reg . ress1 equation of : on (A) Y on X (B) X on y (C) both (D) None ofth . 1· ese I 1. The regress10n mes are perpendicul · each other if r is equal to : ar 10 (A) 0 (C) - 1 .. . . fbe rntn1nusat1on of horizonta~ (13) distances in the scatter diagram, EXERCISE - 5.1 J. Regression analysis is concerned with 10J1 (B) + 1 (D) I. 12. The li~e Yf= 13 -(3/2) Xis the regression equation o : (A) Y on X (B) X on y (C) both (D) None of these. 13. The errors in case of regression equations are : (A) positive (B) negative (C) zero (D) All these. 14. The line Y = a + bX represents the regression equation of : (A) Y on X (B) X on Y (C) both (D) None of these. 15. Two regression lines always intersect at the means. (A) True (B) False (C) both (A) and (B) (D) None of these. I 6. The line X = a + bY represents the regression equation of : (A) Y on X (B) X on Y (C) both (A) and (B) (D) None of these. 17. The regression line of y on x is derived by (A) the rrururrusation of vertical distances in the diagram. ., . 1'be slope of the y is : resreaaion line of x on (A) b (C) 1; (B) b,., (O) . ear equations Y = a + bX and 27, If b ·'>· (D) 111 lib,.... Jll 1 ' bY 'b' is the: Yt and b are . J8• .,, a+ A) · . . ·'>· negative, then r is : ( )(,.. . tercept of the line. Positive (B) . (A) in (C) zero negative pe of the line. 1 28 F h (D) None of these. 0 (J3) ~:th (A) and (B) · xr I e regression equation of Y on X 2 +3y. (C) ' A .,. SO = 0. The value of b is : None of these. ( (D) . ) 213 {B) - 2/3 y., fbe Jine of regr~ss1on passes through (C) - 3/2 19· t1te points, bea~mg ......... number of (D) None of these. 29 s· · mce Blood Pressure . ts of a person on both sides : i potJl dep.:o<ls on age, we need consider : (Al equal (B) unequal (A) th e regression equation of Blood C) zero (D) None of these Pressure on age. ~be slopes of the regression line of y on th (B) e regression equation of age on zo. Xis: Blood Pressure. (A) b,y (B) by., (C) Both (A) and (B) (C) b.« (D) byy. (D) Either (A) or (B). 36· A small value of r indicates only a ....... The equations Y = a + bX and ZJ. X== a + bY are based on the method of hnear type of relationship between the variables. (A) greatest squares (A) good (B) poor (B) least squares (C) maximum (D) highest. (C) both (A) and (B) 31. Feature of Least Square regression lines (D) none of these. is that the sum of the deviations at the 22 . brr is called regression coefficient of : Y's or the X's from their regression (A) x on y (B) y on x lines are zero. (C) both (A) and (B) (A) True (B) False (D) None of these. (C) both (A) and (B) 23, In linear equations Y = a + bX and (D) None of these. X =a+ bY, 'a' is : 32. The square of coefficient of correlation 'r' is called the coefficient of : (A) intercept of the line. (A) determination (B) regression (8) slope. (C) both (A) and (B) (C) both (A) and (B) (D) None of these. (D) None of these. 33. The coefficient of determination is 24. The regression coefficients are zero if r defined by the formula : is equal to : explained variance 2 (A) 2 (B) - 1 (A) r total variance (C) 1 (D) 0. unexplained variance 2 25, bxy is called regression coefficient of : (B) r =I total variance (A} x on y (B) y on x (C) both (A) and (8) (C) both (A) and (B) (D) None of these. (D) None of these. I 1 2' (~ 'CC) bodi(A)and (D) Ndntof 42. If llopa of two equal. tbea rfl (A) I (C, ANSWERS 5. c. ... o ,a. A. 20.8. 28.8. 3'. A 13. o. 21. 8 29. A. 37, A. 0 6. D. 14. A. 22, B. 30, B. 38. B. 7. D 15. A 23. A 31, A 39. C ient1 b.., llbxr • hbuv b IXERCISE - 5.2 . . ,, - , lllmHtS. ofabyilpven • (lc/h) b 11 b • (hlk) b 'YI 11 cbanacd cdtob+k Wrill ti,, co~ 1111nwr : (C) y-y=r(a,lay)(x-i') (0) None of these. 3, The Rpaaon coefficacnt of_ pvea by: (A) , (01 I a,) (B) , (a. (C) (a,' 0 1 ) (D) NOif 4, Tbe two rcgJ'ClllOD ;.aq:..mdlculartocteh (A) r• I (C) ,-o .. Test: Quantitative A titude Stu.., Common Prof1c1enc ~ YL--------~~(A) more thin I (B) less than l D) None of these. ( (C) less than mo • for I b,vanate d1Stnbu11on (x,. Y,), 13 ,. • J. 2• .... n •. hn 1s defined as : rcr, Cov(x,y) (A) (C) (8) - , (crJ er' rx,y, +[(Ex,)(1:y,)lnJ L x,2 -(Ex, )2 I 11 rrx, -.r')(y, -.flJ 2 (D) L(X, -)' -x. 24. lfthe two regression coefficients are 0.8 and 0.2. then coefficient of correla11on r is : (8) - 0.4 (D) None of these. (A) 0.4 (C) 1. 6 JO. If p (x, y) = 0.4 and b,,. = 0.2, th equal to : en b"I:& (A)_ o8 (B) 0.2 · (C) 0.8 . (D) ± 0.8. Out of the two Imes 31 • d of regress,· on D~ by x + 2y = 4 an 2x + 3y _ 5 ,, c,,v~ regression line of x on y is : O, ~ (A) 2x + 3y- 5 = 0 (B) x + 2y = 4 (C) x + 2y = 0 (D) The given Imes can't be regr Imes. ess~ 32. If the correlation coefficient between two variables 1s 1, then the two 1. ~ Ines 0f regression are . (A) parallel (B) perpendicufa 1 (C) coincident (D) None ofth esc 33. Coefficient of correlalion 1s : ion ,Anal sis 15 ' y =80,cr,=2,thea1 ,.1,,I6' JI· 15 , t~en the :sti~t~d value of y p onding to x - 25 is . corresP (B) 140 rf: o. ,,,,. :::: (C) . f,v"' . en the following data : b'>. = 2.33, 0.39, th.en the value of correlation ,~ ffjcient r 1s : coe (A) 0.39 (B) 0.79 (C) 0.95 (D) 0.85. . n the following data : 41 • Give r(x- J)(y- y) = 3900, (x-x) 2 == ,IO, r (A) G.M. offththe coeffiffic~ent of regression. 6360, t (y - y)2 = 2668, then the (B) AM. o e coe 1c1ent of regressi0tt regressIOn coefficient by., is : (C) H.M. of the coefficient of regress~~ (A) 0.5 (8) - 0.5 (A) 0.613 (B) 0.363 (D) Product of G.M. and A.M. of lhc (C) 0.25 (D) None of these. (C) J.363 (D) 2.363 regress10n coefficients. 26. If b-'). = - (413), by, · - (1/3), the value 34. For a b1vanate data, the two lines of · 42 , Given the following data of a bivariate ofr is: regress10n are 4x - Sy + 33 == Oand distribution · b,y = 1.36, byx = 0.613, (8) (2/3) (A) (213) 2x - 9y + 127 = 0. For this data r"" then the coefficient of determination is (D) None of these. (C) (1/9) (A) 2/9 (B) 4/5 given by· 27. Regression equatwn of Y on X is (A) 0.634 (B) 0.834 (C) 5/4 (D) ✓10/6 Bx - IOy + 66 "" 0 and cr, ~ 3. Hence (C) 0. 734 (D) 0.534 35. Slope of the regress10n equation of y 00 Cov(X, Y) is equal to ; XIS. 43. Given the following data for a bivariate (A) 11.25 (B) 7.2 d1stnbut1on . bxy = 0.756, b,y = 0.659, (A) b,> (B) by, (C) 2.4 (D) None of these. then the coefficient of non(C) llb,,., (D) lib,>. 28. The coefficient of correlation (r) and the detemnnat10n 1s given by : 36. If the two Imes of regression for a two regression coefficients byx• bx; are bivariate data coincide, then : (A) 0.502 (B) 0.402 related as : (A) the two variates are independent. (C) 0.602 (D) 0.702 (B) there is a perfect correlation bet• 44. For a bivanate data, the coefficient of (B) r = b<>' x br, ween the two variates. correlation r = 0.91, then the coefficient (C) there is not a perfect positive correla• of alhenation is given by : (C) r = b<>' + br, tion between the two variables. (A) 0.5 12 (B) 0.412 (D) None of these. (C) 0.312 (D) 0.712 (D) r = (Sign by,) Jbxyhyx . 4s. If the coefficient of correlation between 37. If x = IO, ji = 50, crx = 3, cry= IS, 29. The two lines of regression meet at : X and Y is 0.65, then the coefficient of P = 0.9, then the estimated value of x (A) (X, Y) determination is : corresponding to y = 100 is : (8) (a,, er,) (A) 0.48 (A) 19 (B) 20 (C) (a,2, a/) (B) O.Si (D) None of these. (C) 18 (C) 0.42 (D) 21. (D) 0.32 25• If b,. ==- (312), b1 , will be: ... (116), the value of r ~i.~ 150 (D) 145 130 ' (C) the following data : n == 7 · en , 39. oiv ""- 15, r dy =-0.13, 'f.dxdy== 18, i ~ "" o.89. The value of b·'>' is : i 18 67 (B) 19.67 (M . 67 (D) 29.6. 20 (A) (A)Ut (C} 0.74 47 • (B) OM (l))OM ~ 1be foUowaig raallt of 11ae twO lineaor...._tioma..,._dala (X, Y); CJ-'= 2.5, r = -0.8. The mndanl error of estimate of X on Y is : (A) 2.5 (q 1.5 (B) O.S (D) 1.2S 48• Given the followiug bivariale data (..r,y) ; b_., == 415, h,. = 9120, then the coefficient of detennination is given by : (A) 0.26 (B) 0.36 (C) 0.46 (D) 0.56 49, For the following data of a bivariate distribution (x, y) : (a/a) = 2/3, bxy == 315 the coefficient of determination is: (A) 0.9 (B) 0.79 (C) 0.83 (D) 0.81 SO. For the following data; r = 0.6, a,. = 4 the standard error of estimate of y on x is: (A) 0.84 (B) 0.74 (C) 0.64 (D) 0.54 St. Given the following data of a bivariate distribution : hp = 9120, hJ« = 415. Then the coefficient of allienation is : (A) 0.3 (B) 0.2 (C) 0.4 (D) 0.48 S2. For a perfect correlation between the variable x andy, the line of regression is ax + by + c = 0 where a, h, c > 0 ; then p (x,y) = (A) 0 (C) I (B) -1 (D) None.oflbese. .r:. 53, The two Jines or r e ~ x + 2y = 1 and 2x + Y • 1. •~ regression equation of Y OD.~ a. (A) 2x + y = 1 (B) (C) X + 2y = 0 (f)) ~ ~tt?i•"l i Test· Quantitative A p t h ~ . . ... - ~ . en the following data t )( 01 ,1. 2: ; .. 300, 2: XV = 7900, t )(2,..111 ", 650oI and N • 10I 2: y2 -- 10000 I then r is : O8 (B) - 0.8 (A) . (D) 0.6. (C) 0.6 If b . and b . are both positive th 62, )'l .IJ I C!I: I I 2 I I (A)-+-<-:- (B)- +-.~2 common Proficiency M (b + b )/2, then 54- lt!p>-OIJld1tt • "'(8) ffl"' a p ~) "' 2! p (C} '" s p None of these. · are 2.t - 1y + 55. The two lines ofrcgrcss10~ What is the 6 • 0 and 1x - 2y_ + 1 .. . xand I'? Co-lation 11 coefficient between · (D) " (A) - 2/7 (B) 2/7 4/49 (D)Noneofthese. (C) r f regression of Y on X bY Th 56. e me o h following least square method for t e data 5 4 3 2 X 1 II 12 10 y 9 9 is : (A) y (B)X (C) X 8.1 ·t 0.7X 0,718.IY y (D) none of these 57. The statistical method. which helps us to estimate or predict the unknown vulue. of one variable from the known vuluc of the rcluted variubles. is called ; (A) C'orrclntion (B) Scatter diugnnn (C) Regression (D) Dispersion. 58. lf the two lines of regression arc J.\' - I' - 5 0 und 2x y - 4 0, then 'i:' 11 ,;d .I' n:spccllvcly un· : (A) I nnd - 2 (B) · I Ull(I 2 (C) 2 and I (D) - 2 and I. 59, For two v11ri11hlcs .,· und y with sumt' mean, the two regression cquut1011s urc /, y• llX ➔ /,, I +a (A)T;;; l - 11 (C) ~ x • ay 1 13, tht'n ~ is : I ➔ a (B) T;-;; 1- a (D) 1- (I . 60. The regression coefficients of y on x 1s 2/3 end of x on y is 4/3. If the acute angle between the regression lines is 0, then tan 9 is : 2 1 (A) (B) 9 9 1 (C) (D) none of these. 18 brx l7•Y 1 111·1 b.1; .. ' I I r - (D) None of,i...__ (C) -b + -t , <2 ~~ \'. I I) In titting of u regression of' Y on)( 63, .b . . . to bivariate distn u11011, c~ns1s1ing or 1 observations, . the cxplumcd and u:. cxplatncd vunat10ns WCJ'C computed 24 and 3(J, respeclivc Iy. ·1·1le cocfiicicni,, of detcnrnnut1on 1s : (A) 0.4 (B) 0.J (C} o.2 (D) 0.1. M. l'hl' likely production corresponding lo u rainf'ull of 40 ems. from the followi11g clatu : Ruinfitll (111 ems) Oulpul (in quintmlt) Avcrugc : 30 so S. D 5 10 ,. • 0.8 (in cm) ts : (A) 56 (fl) 66 (C) 76 (D) 46.6, 65. The regression cquolions of two vnriablcs X and Y urc us follows : 3X + 2Y 26 O; 6X I Y 31 0. The coefficient of corrclution between Xand Y is: (A) 0,4 (B) 0.5 (C) 0.6 (D) 0.65. 66, For some bivariate duta, the following results were obtained for the two variables x and y : x 53 .2, y "' 27,9, by.r .,, - I •5, b,y - 0.2. The, most probable value of y when x "' 60 1s : (A) 17.7 (B) 16.6 (C) 15.7 (D) 26.7. ion observations Oil' P1f 10 for (.Y), the followina 1 ' ' sllP~~ed (in appropriate unitt 181 ob ~ 13 0, I:y = 220, tr t~2 "' 5506, I: _xy = 3~67: Thhupp t) the price 1s 16 uruta 11 : whel125 04 (B) 26.04 A) . ( 21.04 (D) 15,04, (C) orrelation study of two vattablea 68· In: t the following values are obtained : 811 r ..• 65 I ji • 1 ~ 67, a, .. 2.S, of • 3 s • o I · 0_8. The two regression coefficients r I ) are (b ' 1" (A,) (0.57. 1.12) (8) (1.12,0,57) C) (2. 12, 0.67) (D) None of these. ( . 11 i~ the following information : (jtVC 69· X y Arithmetic mean 6 8 Standard dcviauon ~ 40/3, C ffic 1cnt of corrclatton between X an.ode y is 8/ I5. The. most likely value of y when X • I00 1s : (A) 140.67 (D) 141.67 (C) 241 .(,8 (D) 94.68, 76 The follow1n g cu lculations have been · nude fot prices of 12 stocks (X) on I ' Iiange on a certain Bornbay Stock be day along with the volume of sales in thousands of f; harc (Y). 2:X • S80, l: ,.\'2 41658, 2: Y 370, 2: Yl 17206, 2: XY 11494 From thc~e calculations, !ind the regression equation of prices of stock s, on the volume of' the sales of share~ . (A) Y I. I 02X l 82.33. (B) X 1.102Y I 82.331. ((') X 4.4 l 0.2Y (D) None of these. 71. The following figures relate to the years of service and income in hundreds of rupees of the employees of an organisation. x 8, Y 5, b,x • 0.7S. The initial start for a person applyina for a job after having served in another factory for a period of 12 years in a similar capacity is : I w bet ~ be most lilt, heiaht" (A)IOJ5 (C) 73 7 74, Oivenx llne1 of re re1pectivety, If M of correladc,n (A) 0.4 (C) 0.6 75, Which otthe fol true (A) The rep applyma tbl lq\llrll, (8) TbeCOffela (C) b11 1114 b potJtlvo m (D) Tbo,.1ipt0 b.,or 76, You~ '"' cortt fa 0, (~ . . . obtained 9f' dafl ,!D tw0 Y· f •20, y • 15, . ot X • 4, Standard ~y • ); , • the coefficient of • o,7. The likely value of Y, •Uu: (.l) 17.9 (B) 17.8 (C) 17.1 (D) 16.1. 71. In a partially destroyed la~oratory record of an analysis of correlat10~ data, only the following results are legible : Variance ofX = 9 Regression equations : 8X-10Y+66=0 ... (1) 40X - 18Y = 214 ... (2) On the basis of the above information, the value of cry is : (8)3 (A) 2 (D) 5. !C) 4 79. The following data, based on 4~0 students, are given for marks m Statistics ~nd Economics at a certain examination. Mean marks in Statistics = 40 Mean mars in Economic = 8 S.D. of marks (Statistics) = 12 Variance of marks (Economics)= 256. Sum of the products of deviations of marks from their respective mean = 42075. The average marks in Economics of candidates who obtained 50 marks in Statistics is : (A) 54.5 (8) 45 (C) 52 (D) 49. 80. From the following data : cr.r = 3. b.r, = 0.85 and by.r = 0.89, the value of cry is : (A) 3.57 (8) 3.07 (C) 3.97 (D) 2.07. 81. In order to fmd correlation coefficient between two variables X and Y from 12 pairs of observations, the following calcu!Jtions were made : I X • 30 I X2 · 670, I Y = 5, I Y2 = 285, I XY = 344. On subsequent varification. I discovered that the pair (X = 11, y ~ was copied wrongly, the corre~ ) being (X = 10, Y .= I~). The correlation coefficient 1s : ' (A) 0.61 (B) 0.71 (C) 0.81 (D) 0,67. cor:~ 82. The following ~ta ~hows the nutnber motor registrations m a certain tc . Of for a term of 5 years and the s~~ motor tyres by a finn in that terntoa c or . d. ryf0r the same peno Take dx = (X - X) = (X - 700), dy = (Y - Y) = Y - 1300, then r. dx = 8, r. dy = 0, >.: dx 2 = 27,Soo r. dy2 = 8500; >.: dxdy = 41500 ',,,_ · •oe estimated sale of tyres when the mo . . 1s . 850 1s . : tor registration (A) 1424 (B) 1324 (C) 1524 (D) 1624. 83. The following calculations have been made for pnces of 12 stocks (X) 0 . n Stock Exchange on a certam day alon with the volume of sales in thousands 0~ shares (Y). From these calculations the regression coefficient of prices of stocks, on the volume of sales of shares is : r. X = 580, >.: Y = 370, >.: XY = 11494 r. x 2 = 41658, r. v 2 = 11206. · (A) - 1.10 {B) - 2.10 (C) - 3.10 (D) - 0.10. 84. For 100 students of a class, the regression equation of marks statistics (X) on the marks m Economics (Y) is 3Y - 5X + 180 = 0. The mean marks in Economics is 50 and variance of marks in Statistics is 4/9 of the variance of the marks in Econonucs. The mean marks in Statistics is : (A) 56 (B) 55 (C) 66 (D) 65. 85. A departmental store gives in service training to its salesman which is followed by a test. The management is considering whether it should terminate ,ervicos of the sa lfle well in the test. The dO tes to the test s~ ;~e by the salesman : "' ,:::. 20. y = 40, .% ... ;c y _ y ; t d x2 = 120, t f d.t r dY::::- 11 ""9, the correlation between the til.1.· 1 . ~t.and the sa es 1s : score . 5 (B) 0.85 (A ) - 0.9 (C) _ o. 75 (0) 0.9S. some bivariate data, the folloWing 86· for Its were obtained; Mean of Variable resu X"" 53.2 and of Y = 39.5; Regression a-:cient of Yon X = - 1.5 and ofX coe111 Y on ""_ 0.38. What should be the most likely value of X when Y = 50 ? (A) 39.21 (B) 49.21 (C) 48.21 (D) 59.21. The lines of regression of a bivariate 87, distribution . are as 10 "' IIows : 5X- 145 = - IOY; 14Y - 208 = - 8X. fben 5 The mean values (X, Y) is : (A) (5, 12) (C) ( 12, 3) S8, While calculating the coefficient of correlation between two variables X and y the following results were obtained : The number of observations N = 25, E X = 125, I: Y = 100, :E X2 = 650, E Y2 = 460, >.: XY = 508. It was however, later discovered at the time of checking that two pairs of observations (X, Y) were copied (6, 14) and (8, 6) while the correct values were (8, 12) and (6, 8) respectively. The correct value of by.r is : (A) 0.6 (B) 0.7 (C) 0.8 (D) 0.65. 89. Given : Unexplained variation= 19.22, explained variation = 19.70, then the coefficient of correlation, is : (A)± 0.71 (C) ± 0.61 t (B) (12, 5) (D) (3, 12). (B) :!: 0.75 (D) :c 0.65, If ~) (C) 95. There and y fi Ix=SS, ty2.= 11 (A) 1 (Gl2 "· eommon proficiency reel: Quantitative Aptit~'- • (A) 4r ti. la I panfally desiroyed l1bor11ory record of ■n 1nalysi1 of • correl11i~n data. the only information remained is ••-•lo l --..-..- n equal on• : • - 9 - I07 "' 0. 4.r Sy + 33 0, 2x Y Then lhe means _ values of x and Y are : x • 13, y • I 7 (A) (8) - 12 • • 13 ' ~ (C) ;r .. 15, )' a 13 (D) None of these. : 99. For the following data, x = 36, )' -= 85, a.• 11. a, = 8, r = 0.66, the regression line y on x is : (A) y = 0.48x - 67.72 (B) y = 0.48x + 67. 72 (C) y = 0.91y- 41.14 (D) None of these. 100. If the lines of regression in a bivariate distribution are given by x + 2y = 5 and 2x + 3y = 8. then the coefficient of correlation is : (B) 0.866 (A) - 0.866 (D) 0.666. (C) - 0.666 101. 4x + 2y = 1 and 2x + 3y = 4 are two lines of regression of x on y and y on x respectively. Then the coefficient of correlation is : -11 ✓3 (A) 1/ ✓ 3 (B) (C) 11 ✓ 2 (D) -11 ✓ 2. 102. For the following data : L x = 110, Ly = 70, L x2 = 2500, L XJI = I00 and n = 20, estimated value of x when y = 4 is : (A) 5.42 (C) 5.3 2 (B) 5.62 (D) none of these. 103• Let X and Y be two variables with correlation coefficient r. If the values of X and y series are changed such that the Cov (X, Y) remains unchanged while ~e variances of X and y become 4 hmes ~eir original values, then the correlation coefficient between X and y becomes: (B) (r/4) ~ l6'' (D) (C) . (r/16). 104. If the coe~c1cnt of cor_relation b X and Y 1s 0.28, covariance b cti,,e... • 6 d h ct.., -ti and y is 7. an t e variance o ten~ then the S.D. of Y series is . f )( ~ g (B) 9.J · ' (A) 9.05 (C) 9.08 (D) 10 .Os. 105. If the two lines of regress 10 x + 4y '" 3 and 3x + y == 15 n arc th value of x for y = 3 1s : ' en !ht (A) - 4 (B) 4 (C) - 3 (D) 3. J06. If 0 1s the angle between th regression lines with e ~o correla1· coefficient r, then ion (A) sin 0 ~ 1 - r2 (B) sin 0 :s; ,.i (C) sin 0 ~ r2 + I (D) sin 0 > I - I - -~ 107. Let x = 15, Y = 80, cr. == 12 · r ' (J "' 12 r = 0.75. Then estimated valu , y correspondmg to x = 55 15 : e of (A) 110 (B) 120 (C) JOO (D) none of thes 108. Two random variables hav e. th regression Imes 3x + 2y == 2e6 c _ and 6x +I [ -b 31 . The coefficient of corre at10n etween x and y is given b . (A) - 0.5 (B) 0.5 y. (C) 0.25 (D) None of these 109. The lines of regression of y on X and X on Y make angles 30° and 60, respectively with the positive direcn ~ . of x-axis, then the correlation coefficient between X and Y is : (A) 1 (B) - 1 (C) (11 ✓3) (D) ✓ 3. 110. The correlation coefficient when the line of regression are 2x - 9y + 6 = Oand X - 2y + 1 = 0 is ; (A) 2/3 (B) _ 2/3 (C) ± 2/3 (D) None of these. 111. 3x + 4y - 7 == 0 and 4x + y - 5 = 0 are the equations of two regression 1ine1. The correlation coefficient between x and y is : f.) o.43 (B) - 0.43 (D) -0.34. (c, o.34 ( between the lines of regrCllion 1J. Jf r rid y on x, is± l, then the: Ob I oriY . 'd Jjoes comc1 e (Al lines are perpend'1cular (B ) lines are not parallel (C ) J'lone of these. (D ) and cr are standard deviations f Jf (Jr. \ 0 X . 113· ao d y· senes and r the correlabon y at (A)_!_ 1-a 2 coefficient, then cr x- .l (i.e., Variance of t _ y) equals : · (C) cr/ + cr/ + 2r cr. cry a,2 + cr/ - 2r cr.r aY 11 (a/- cr/ + 2r s, cr) (D) none of these. (A) (B) If u = x + y, v = x - y and x and Y are 11 4. independent then cr.-' 1s . equal to : (Al a/ (C) cr/ - cr/ (B) cr/ + cr/ (D) cr,2. JtS. If x and y are related by y = mx + c, m and c being constants, then coefficient of correlation r between them is : (B) 0 · (A) 1 (D) none of these. (C) 2 !16, If the values ofx; and Y; are transfonned to u, and v, by the relations u; = ax, + b, v; = cy; + d, then which of the following is not true ? (A) cr. = acr, (B) cry= cay (C) Cov (u, v) = 2c Cov (x, y) (D) p (u, v) * p (x, y). 117. Angle between two lines of regression is given by: (C) -1.. 1-• t, (D) - . 119. . 1 ex The regresnon equations af'y mt-.r l1id x on Y are respectwely Y <Oil x ad 4x - Y = 3, then the correlation coefficient between x and y is : (A) 0.5 (B) - 0.5 (C) 0.7 (D) - 0.7. l20. The standard error of estimate of X on Y is given by: (A) cr.r (1- r2)112 (B) a_. (1 _ r2) (C) cr, (1 _ r2)112 (D) 01 (1 _ r2). 121 • The Cov (x, y) between x andy for the following data is : x: 3 4 5 6 7 y: 8 7 6 5 8 is: (A) 2 (C) 3 (B) - 2 (D) 4. 122. The coefficient of correlation between x and y for the following data x: 1 3 4 S 1 y: 3 7 9 11 15 is : (A) 2 (C) l (B) 3 (D) 0. 123. The regression coefficient b.\)· for the following data : {xy: (5, 2), (7, 4), (8, 3), (4, (6, 4)} ,>.. is: (A) 0.5 (C) 0.7 124. The regtession following {(.xy) : (1, 6 is: ( ~ Common Profic1enc S.21 (B) 0 (A) 014 (B) 0.64 (D) 0.74. . (C) 0.36 . of determination . the coefficient 157. G1ven • . = 0 9025 the value of r is . . ' 0 95 (A) 0.85 (B) . (D) 0.65. (C) 0.75 . data : b ' == 0.39, . !'- o9025 158 Given the following • coefficient of dete~ · atIOn IS • ' then the value of b_ry is : (B) 2.314 (A) 3.314 14 (D) 4.314. (C) OJ · tion is 159. If the coefficient ofnon-de~e ~ 0.502, then the value of r is . (A) 0.69 (B) 0.71 (C) 0.73 (D) 0·8 1. . . . of non-detennination 160 If the coefficient . . 1s . 0.502, then the coefficient of alhenat1on is (A) 0.71 (B) 0.61 (C) 0_51 (D) 0.81: . . 161. Given the coefficient of alhenatto~ is 0.4, then the coefficient of correlatton is : (A) 0.91 7 (B) 0.817 (C) 0.617 (D) 0.717. 162. Given r == 0.4, cry = 0.6, then ~he standard error of estimate of Y on X IS : (A) 0.45 (B) 0.65 (C) 0.55 (D) 0.35. l63. Given cr, == 4, r = 1, then the standard error of estimate of X on Y is : (A) 1 (D) 3. (C) 2 Given byx = 3, CJ>= 1!, and r:::, 1 th 164 · tandard error of estunate of)( ct\ 11. s 011 y ., (A) 4 (B) 3 is : (C) 2 (D) 0. 2. B. 10. A 17. A. 18. C. 25. C. 33. A. 26. B. 34. D. 42. B. 50. C. 58. A. 66. A. 74. B. 41. A. 49. D. 57. C. 65. B. 73. B. 3. B. 11. B. 19. A. 27. B. 35. B. 43. A 51. C. 59. C. 67, A. 75. B. 4. C. 12. C. 20. B. 28. D 36. B. 44. B. 52, B. 60. C. 68. B. 76. C. 91· s 13. ,o , 13. 113· 13. 1il- 13. ,i9· p. 13" 3 ;.. 14 . (D) none of th . wh1c . hi . a ':ays r1es on theese• 166. The pomt lines of regression 1s : liiio 167. If the tw~ l!nes_ of r~gr~ssion of bivariate d1stnbulion coincide, the a correlation coefficient p satisfies n the (A) p = 0 (B) P > 0 (C) p < 0 (D) P "' 1 or _ I 168. The regression line of Y on X i y = 3x + 10 and that of X on y is s X 1 3 10 = - y - - · If (Jx = 4, then 3 (A) 3 CJ :: >' (B) 48 (C) 12 (D) none of these. 169. In a bivanate data :Ex = 30. :E y "'400 :E x2 = J96, :E xy = 850 and n == 10. Th~ regression coefficient of Y on X is : (A) - 3.3 (B) - 3.2 (C) - 3.5 (D) - 3.4 1 ,s j. ;.. 1S9• 13. 1 v. 16 • 82, C. 90. D. 98. A. 106, A. 114, A. 122. c. 130, C. 136, C. 144, D. 152, A. 160, A. 168, C. 5. A. 13. C. 21. C. 29. A. 37. A. 45. C. 53. B. 61. A. 69. B 77. C. 6. D. 14. C. 22. C 30. c. 38. B. 46. B. 54, A. 62. B. 70. B. 78. c. 7. B. 15. C. 23. B. 31. A. 39. B. 47. C 55. B. 63. A. 71. A. 79. A. 8. A. 16, B. 24. A. 32. C. 40. C. 48. B. 56. A. 64. B. 72. A, 80. B. 83, A. 91. C 99. B. 107. A. 11S. A. 123. D. 131. A. 137. D. 14S. C. 153. C. 161. A. 169. C. 84. C. 92, C. 100. A. 108. A. 116. D. 124. D. 132. A. 138, B. 146. D. 1S4. C 162. C. 85, D. 16. a 93, A. 101. B. fl, B. 95. -A. 103. B 111. B 119. A. 127. A. 133, D. 141. D. 149. B. 157. B. 165. A. 94.e, 102. A. 110. A. 118,A. 126, B. 132. D. 140. B. 148. A. 156. B. 164, D. 109, C. 117. B. 125, A. 132, D. 139, A. 147. B. 155. C. 163. B. & &a ~A. *· A. tu.~120. A. 128. B. 134. B. 142, B. 150. C. 158. B. 166. A. SOLUTIONS/H1NTs ee Text. t,.)' S l( ): see Text. z(JJ : See Text 3(1J)· see Text 4(C): e Text 5(;\): Se . If x and Y are two independent 6(0)· variables, then the correlation oefficient r between them is zero. C • • Putting r = 0 m equations of the two Jines of regression, we get Y = Y and X = X which are of the type x = const. and y = const. byx + bxy >r 7(B): We have to prove that ~ Or ANSWERS 1. A. 9. A A· t}· A· (C) 4.71 (B) (b_ty, by,) (D) (0, 0). ssiO 13· sl• 165. If r = 0.5, b.'9' = 0.6 a~d cry::: 2.6, the ndard error of estimate of)( tq~ 0 l\y . sta (A) 2.71 (B) 3.7} Is. (A) (x, y) (C) (crx, cr) ~ ~ ) (ray l a x)+(ra x lay) --'--------->r 2 cry2 + crX2 > 2crX crY or cr/ + cr/ - 2crxcry > 0 ⇒ (ay- aJ2 > 0 which Is true. S(B): The lines of regressio~ of Y on X and X on Y are respectively a1x + b,y + c1 = 0 and azr + b-Jl + C2 = 0. Therefore, byx = slope of the line of regression of Yon X =- (a/b,) ...(l) (1/bxv) = slope of the line of regression of X on Y = -(a/bJ x-a byx = (rcr/cr) = r · (kcrjhcr.,) = (klh) rcr/ o.) = (lc/h) b,,,.. lO(A): The regression coefficients are independent of origin but not of scale. ll(B): Here m1 = 2/5, m2 = 8/5 ffl1 - ffl1 :. tan 8 = 1 + m1m1 (8 / 5)-(2 / 5) 30 = 1+(8 / 5)(2 / 5) = 41 12(C): See Text. 13(B): Since G. M. > H.M. :. G.M. of (byx and bzy) > H.M. of (byx and b..,) ⇒ r> !. 2byx · bxy ⇒ byx + bxy > b bxy + bxy by_.. xy r 1 1 2 ⇒ -+->-· b"' b"' r 14(C): bzy = r (a/a,) and by,= r (a/az>• . bzy x bYJ = r (a/a) .. ., x , (a/aJ •r. :. bxy = - (b/ aJ Since bvx · bxv ~ 1, therefore y-b v = - - where a, b, h ' k .h and k are constants. 9(A): Let u = - tS(C): See Text. Ion llnts .,· ,m J' and .I' TIit two N1rffl ar rhrlr nit•ans on x always lnrtrst'C I (.r,)'), ~ • ,. or line or =- ,.~ 41( s 41( :S I ( .'. 0 :i , J :S I) 41( 1•,y ....:d) .;;;:;:/I-:; ~ b,, x 0.2 . (0.4)i => h.t'' 0.H. 23(8 ): Sec Text. 0 8 x 0,2 • 1 6 => r • 0.4. 26(8): r • I J2(C): Since ,J I ⇒ Cov( x, y) = a/ i 5 9 X 4 36 Cov (x, y) =- - - .., 5 5 1.2, 2 28(0): p = byxbry. But br:r, h,Y' p Cov (X, Y) are all of same sign. :. 29(A): p = Sgn hrx Jbyxbx, • (X, Y) , by definition. JO(C): We know that b,,Jxy = p2 J ' 0 b11 h1.,, when,. 80 , X h,, .. 33(A): Smee ,:i (,, =r- I II -:-._ I, '> 18 _'( 15)( O.IJ) 7 89 b1·1, . , r G.M. of lht• cocfflclll rcgrc,slon. 11h or 34(1>): Ilcrc, I\, 2/9 ond 11/,,y slope or 4.1 - Sy., 33 • 0 I 4 5 or - • h,v • /Jn, 4 18 0,279 11 ,o,n: p• 4 JJO .!:(x (Note that p , 0 as both /, and 11 •·t 11.arc posillve) · 35(B): Regression equation of y on x 18 y y hv., (x x) Its slope by.,· · 36(8): The two lines of regression coincide If there Is a perfect (direct or Indirect) correla tlon between the two variable~. 27 9 • ~m • 19,97. 21 0,9 ( ~) 1 0.18. 150 50 Hence, regression equation of X on Y jg: X 10 0.18 (y 50). Fory= 100,x 10 + ,18 (100-50) =- 19. )< x)(y-y) l.(x o.•w • 0,897 O.9S. -2 x) 3900 ·6360 - 0.613. 42 rl • 43 cM: s •r• 0,834, o 1s6 x o.6s9 ('oclflclcnt of ll<JU-dctermh111tlon (k 2) J r2 I 0.498 0.502. 44(B): Coefllclent of ulllcnatlon (k) ,. (0.,,1. o.a1. SO{C): 11nd•rd Error eatimate of y on x 11 given : s,... a, (I - r)llt 51 C: 4 )( ( I 0.6)2 4 )( (0.4)Z - 0.64. • ( ): ,i h,y >< br, (9/20) >< (4./S) • (9/25). Coernc1ent or Alllenatlon • (i) • (I (I 9/2S>''z • (16/25)' 12 4/S 0.4. 5 l(R): The !incur rclation1hip between the lw<, voriuhlcs 1s "x ~ by I ,. 0 with alopc - !!. < 0. . b So, t~c two variate~ arc in perfect negative correlation und hence p (X, Y) • -1. 53 (H): If firnt line x 1 2y 7 is taken as rcgrcsswn equation of Y on X, then Wt wrilt it a~ : I Ji-{0.91) 2 • .ft- 0.83 • Ji 7 ., 0.412. 4S(C): Coefflclent of determination r2 (0.65) 2 0.42. 46(8): Coefficient of non-determination 1 - r2 I (0.25) 2 0.063 0.94, 47(C): Standard error of estimate of X on Y b )'I 2 Al~u, the Becond equation (which will ht rtgrc~sion equation of X on Y) can ht written a~ I 7 x• --y+2 2 -_, hty I --. 2 1 1 So,hrAr• (~ )(~ )-¾st. Hence regre11lon equation or Y on X lsx+ 2y • 7. 54(A): Since p > 0, therefore by, and b.ry ari: also positive. byx +hxy fi:'"T Hence - -- ~ ,.;byxhxy 2 ( ·: A.M. ~ hyx +hxy G --i-t?vfr ~ ~ c!! p s. . is given by : y a,.. (l - r)i/l 2.5 (1 0.64) 112 • 2.5 >< 0.6"' l.!, 7 --x+::::) 2 2 0498 .. ✓(1-r)2 u,. 1 3 09 5 2 " Coefficient or determlnadon y bry X hy,, r ",y x h,.t ~--,><3 5 ,.!)(! (,i}: ('ocfflclcnl of d_t·lcrml1111tlon 1.36 x (t613 6 . 37(A): Here, bxy = p ( :·~] h,,, >'. Jo.k97 ., r 5= 10 36 ,.2 • "•Y • r • •,, - •,. • (415) (9120). 49(1)): b • , !.._ ., 0 , r>"1 . ( () IJ)l 7 17,72 1 89 00241 48(8), Cotmc..tor~ ~ 140. l:,(v2 _ ~dy) h 2 5 15) >:.dw(v- (l:tlt) (l:dy) (Y - Y), -x- 15), C,(.\ 2!11, y · HO I 6(25 .... j I, I, 0.75)( 16 ~ I i,or x 2,l' ...... 2 H• Ii II' I 2 6, rciir,·. regression equation of y 1 on X ,s I' S ..... Rt'gtl·~smn lim·s arc y y hl'I (:\ X) ⇒ ( •: r has same sign as of b,r or b>•x) 27(8): Given regression line is : Bx - IOy + 66 = 0 8 66 4 33 ⇒ y= -x+- ⇒ y=-x+10 JO 5 5 '" s und /, ., 9 = J(-4l3)x(-(l/3) "- (2/3). i 5 II C,c, Rt•gn·s,1011 l'ct1111t1 011 of\' 2 l.,· I· Jy ~ 0. . 1111J'li: 0.25. ,p;;;x h,,x b = J\' ,\ 2S(C): r • Jbx, x hp - J(3/2)X (l/6) (1/4) f Thl'Y un• ,•ohll'ld,•111. 22(8): If one of the regression rocfflclcnts ls greater 1111111 unity. then the other Is less than unit)', ?- • b,, x h,, 0,8 111t• --2 ·f }. und .1 - ,t8(p): .·. und X - :::, 0SK!i(l/4), ll(('): r • linl:s g1\'l'11 Clcndy h11 (),bl, 0.2 + 2y 4 llllll l1 .Jo.s X (),4(l h,, ,. JI(,\): The (OA)l ""' -""' i.1•, .'' s I,,pc 20(8): We h11,·e rtgrcssion ofY 011 X I(: :md (_lll:w) sll1pc: ,if line i,r rrgn·ssi1111 lit :,; on y ( !/4). Thnd'im·, ,,,, ;, ~ ---- p2 I 17(A): Sff Texr. 18(C): Sff Tell, 19(A): r .. .fl,., x b,,. 24(A): r~i ~111,, uo bvx + hxv 2 => "' ~ P• ;? G.M.) ---------------:~~cv'1ei!~~~~~~jj~~=--,~1~0~:;::::.-,~~-:-:;:;::::~::~:~~~~i~=-~ ~ · ·:~~~~~-~:!:!=f.S¼: (sincewe bsign ofb)'X must be negative,l = =t ion are I 1be ofresress ... :e regression equation of yon Q .. .<)' twO Hid li- 1y + 6 • 0 5'(11>: ., (1) ... (2) 2t1+ 1 • 2, . and ,x - -; the regression If we take <I) x•'then (2) is that of · ofY on , cquabon W an write these as : X on Y. cc 2 I 2 6 .. -y-,. -x+- andx 7 > 7 7 7 respectively. 2 2 d h•J "' -7 b=-an )'T 7 4 2 2 b -x-=--<1 ⇒ bP .o,. 7 7 49 4 Now, P2bb-- " •Y - 49 ⇒ p Total 9 18 tu 30 48 2 = -7 yi 81 81 4 9 100 144 16 12 121 25 55 II 5 IS SI 160 SS 527 l:Y "'Na + h'f.X, l:XY = al:X + b"i.X2 51 =Sa+ !Sb and 160..., !Sa+ 55h Solving, we get: a 8.1. b - 0.? Regression equation : is Y "" a + hX or Y 8.1 + 0.7X S7(C): Regression S8(A): Solving the equations 3x - y - 5 = 0 and 2x-y-4 = 0, we get: x= 1,y =-2. S9(C): Since y = ax + b and x ay + P, 111 .ai :. y = ax+ b, x But x = y. 00 1-a 60(C): by, • 2 b ii x b,,.. = ~ = + 0.8. G.M > H.M. G.M. of bJ., and b." > l-l i , . .• and . ·••·· Of b .•, b_"<Y 2 byxbx_v r b b ⇒ r > b ⇒ - > -:--!.:...-1J> b1•x + xy 2 bYx +b' b byx + xy 2 1 1 ~ ⇒ --~>- ⇒ - + - 2 bvxbxy r b,-x b,,x >-.. 63(A): Coefficient of deternunation ., Explained Vanation r'- "" Total variation 64(B): Regression of Yon X : = + bxy ,_aY = 0.8 x _3.5 = 1.12; a, 2.5 a.r 0 8 2.5 b = r - = • x- = 0.57. xr aY 3.5 aY 8 40 u b - r- = -x69(1J): nere >•x - ax 15 3 x 5 68(B) byx = 1.422. Regression hne Y on X is Y - 8 = 1. 422 (X- 6) O' A When X = 40 cm, then y - 1.6 (40) + 2 66 cm. 65(8): Solve the given hnes to get X . 4, y 7. Assume 3X 2Y 26 0 as the regression line of Y on X. p, 3 3 ⇒ y - -X 2 + 13 ⇒ byx = --2 . Similarly, from the second line, we 70(B): b,y -6 .... - 0.5. ljl..,.,,._~~.;;.:;...;;, Sitrilarty, b .. -10-0 • 33-(-1)1 /8 • -0.34. ,-~ == ✓(-0.278) X (-0.304) =-0.29. 7 3(1): Regression equation of y on X : o, Y -Y == r-(X-X) o.. ⇒ y - 80 = 0.6 Y = la + y ⇒ bx,, == k. Now r2 = bx, x bJIJC = 4k. 1 If k = , then the result r2 = 4k 16 l 1 ⇒ r2 = or r = ± = ± 0.5. 4 5 0.75 (X - 8) ⇒ Y = 0.75 ~ -1. When X = 12, then Y - 0.75 x 12 - 1 = SorRs.800. 2 As both b..~, and byx arc positive so r must be positive, r = 0.5. 75(B): ·: r = ..}b.., x b1" , the sign of r will be as that of b.., 76(C): Regression equation X on Y : - ax (Y - -Y) X - X = Y- o,, ⇒ X - 36 = 0.66 X 11 8 (Y - 85) ⇒ X = 0.9075Y - 77.1375 ••• (1) Put Y = 75 in (1) to get X = 26.93. 77(C): (Y - Y by., (X - X) Y (X-165) X == 170 cm, then Y = 83.75 kg. 74<B): x = 4y + 5 ⇒ b = 4 and .:ey 17236-12(30.833)2 :. Regre~sion line of X on Y: X 48.333 ( 1.102) = (Y - 30.833) ⇒ X = - 1.102 Y + 82.331. 7l(A) Regression line of Y on X : 10 8 X ⇒ Y == 0,75X - 43.75 When 11494 - 12 (48.333 X 30.833) = - 1.102. Y Ji,;xbx,• ""Yl -N(Y)l ~ = ⇒ lb,.b,, - 11 byx 3467 = 130a + 2288b ⇒ a= 8.8 and b"' t.015. 'fhe line is y = 8.8 + l.015x. put x = 16 to get : y = 25.04. ⇒ y = 1.422 X - 0.532 When X = 100, then Y = 141.67, :EXY-NXY I ... 8 • :::: 107.7 - 1.5x or y Wbell x == 60 , then J' "" 107.7 - 1. 5 x_ 60 : 17,7, rhe line Y on x_ is Y - a + bx, where a d b are given by the nonnai 61(.-\): ::uation 220 = 10a + 130b, and 24 ⇒ Y 50 - l.6(X 30) ⇒ Y 16 X + 2. r= <0 xis: 60 "' 0,4, Y _ Y ,. _ a y (X _ X) :. r _ y "" b>" (x - x) ~ y- 27.9 = (- 1.5)(x- 53.2) ll have b.ty ' - -. 6 3 b.ry "' 3 6 "' o.4 1 4 ,;_.-\)· a. y +p. i(l - a) = b, x(I - a.) = 1-a • , = Jbxy < 0, . 62(8): Since (Note that b.. > 0), so P > O y XY x2 S6(A): X 9 9 I 2 3 4 Similarly, 1ox 7900 - 250 x 300 b .___ xy = 10 x 10000- (300)2 :. So our choice is valid. ' (8/9)-1 6/ 3 18 · NIXY -(tX) (tV) 1 6l(A): b,,x = N(tX ) - ( t ~ 10 X 7900 - 250 x 30o = ~ 10 x 6500 - (250)2 "'l, ⇒ Y) Oy = r- o .. - (X - X) · 0.7x3 Y-15= -4-(X-20) = 0.525 (X - 20) or Y • 0.525X + 4.5, When X •24, so or Hence R ~ t,91 t,.. 100 Cerreded J: 66. Jt: 6SO 193 9x(l20 9) IHI Jf l e,•Jrl =J346 " 9 • t..,• Eq _ •o/ . ,1 • ~r• Jt,r 193 9x(346 9) • x bq 193 12 R .._ 330 or lt(D): a. 180 ( ⇒ r2 I J120xk6 = 203 Cerrected % S08 6xl 520. 96(A): The regression line : Y - Y • b,, (x _ x) or y- 8.8 "" 1.24 (x - 5.5) ⇒ y• J.Z4x+ 1.98. 97(A): 4x + ~1• - I = 0 ⇒ y (-- 416) ⇒ blt). = - (2/3). 4.x - Sy + 33 2.x - 9y - 107 = 0, we get ; x "' 13, y "' 11. 98(A): Solving ray 99(8): b_..,, t03(B): Herc r :::: X = 0 and (2<Jx)(2cry) - ~ 1 X Oy '", 4t, 104(A) r = ⇒ 0.28 7,6 <1x <1y ~ cry =9.05. (J _ = ,2 )2 + ,2 (~]2 II r2 = hxy / + ,2 x = (I 10/20) = 5.5, y = (70/20) = 3.5 II [crt +~] sin 2 8 s;; (1 - r2)2 ⇒ sin 8 s;; 1 - r. ⇒ / 107(A): hxy =r(aylax) = 0.75 X (12/12) = (3/4). Regression line y on x is : ⇒ y= ~ hyx X ⇒ bxy I r = - (1/4). [(y-y)2 ±2(x-i)(y-ji)) a,.2 • a,+ay±2ro 2 2 a ~02_2 X ( J/4) (3/16) = - 0.43 nearly. I in 1-,2 Ou::o2+ 2 x o,, +2r CJ• CJ and 2 , • -o, +oy -2rc, c, N • ow: and Y are independent, so r• 0 y ... C1 u = 02J( + c,2y Also r = (IIN)l:(x-x)(y-ji) [✓(IIN)F.(x-x)2 ° --X--Y-, r 0 2 +02 tan 8 = 0 => 8 =o ✓(II N) (f.y- ji)2] :::) r= (IIN)mI:(x - x')2 (II N) Jm 2 (l: (x _ x) 2 ±m ± I :. u-u ;;;. a(x-x) ... (I) Similarly, v - v = c (y _ y) ...(I) :. a; =(II n) 2 = a a;. r. (u -'ii)2 =(1/n)I: a 2 (x-x) 2 (using (1)) Similarly, cr! .. (1/n)l:(v- v,2 a x CJ X •y o2- 2 = - ( .jj /4) = - (1.732/4) =± 2 ,. -o, + oy -2, ax a,. 2 m Queatlon No. 113, we have 114(A): Fro 116(0): u = ax + h ~ 'ii= ax + b. x bxy }'X ll2(A): Putting r tan 0 ~ = Since r, h,y and b,y have same sign :. 0/N) t (X - m (l/4)y+(5/4) =b X)2 • X (3/4). =± (✓314) :. r • (x - ?) :I: (y - Ji) (II N).t(w - w)2 (3/4)x+c114) =- (3/4) ⇒ y- 80 = (3/4) (x 15). Putting x = 55 in it, we get : y = JOO. 108(A): The line 3x + 2y = 26 can be written )' = (-3/2) X + (26/3). ,., (l) The !me 6x + y = 31 can be written as x = - (116) y + (3116). ...(2) Now byx = - 3/2 and b.IJ' = - 1/6. :. r2 = (- 3/2) (-1/6) = - (1/4) = - 0.25 ⇒ r = - 0.5 (·:bzy<O,so r <0). / - - W• (x:l:y)-(f :1:Ji) :. Y - ji = m (x - x). Equati~n of the line of regression of x on y 1s : 4x + y - 5 O ~ 141 l 15(A)·. y. mx+c::) Y=mx+c t1HB): Equati~n of the ltne of regression of y on x 1s : 3x + 4y - 7 = o Now r2 I o. 64 hYIC - (219) (219) r = (2/3) as byx > 0. y-y=bxy (x-x) 20 X 100 - I 10 X 70 = 20 X 2000-(70) 2 =- 0.1 6• X ~ r = ± .J4;9 = ± (213) Ox Oy Regression line X on Y is : X - 5.5 = - 0.16 (Y- 3.5) ⇒ X= 6.06 - 0.16Y. When Y = 4, then X = 6.06 _ = 5.42. xy X Oy ~ (I_ ,2)2 _ (219) x + (6/9) )' _. (2/9) X + (2/3) Equation of the line of regre ssion . of 2y I Regression coefficient of . b = 2. x ony is x cry sm 2 0 (I _ ,2 )2 nl:/-(I:/) Also, 0 {1/ ..fj ). x on y ls : x cr 2 (l-,2)2 Now . Regression coefficient of Is hyx "' (2/9). y on x I [at + ]2 --t ,2 (I-,2)2 = - (2/3). I02(A): b = n.Exy-(..tx)(I: y) i e. cosec 2 0 = I + co12 8 = 1+ + (4/3) (sign of r 1s to be taken negative as both b-cy and byx are negative). or y CJ X :. ... iii•x :t Y b>•x (·: r > 0 as b •Y > 0) )' equation Of the line or regres 110(.4 · .Y on x Is : 9y 2x + 6 Slon or .I 0 X ::= ((II ✓ 3) X (1/ ✓3))'12 ) 1-,2) tan0 =± (- , - ~ + a: ⇒ =-(1/ ✓3). ), I.et" . :. r:::::. Jbxy 105(B): Putting y - 3 in the line of x on y, i.e., 3x+y 15 w regressj , e get, o~ 106(A): We know ·-t"'4 Now r =Jbxyx hyx xy 1~ Cov (X, Y) ~ ⇒ ~'.n10: sJopc of regression line .,., on .t , ~ ---.__ Ht ,"'' hyx • tan Joo {I/ .,/j ' ore •• ' ) sJope of regression line .t on Y I IJ(B . .._, are ; , : : ~• coblddua or "" (J lb,y) = tan 600 "' ,fj X+u "' .,,u • x-yancf • x :l:y ::J b,y (1/ ✓ 3). 4 , __ Cov (X, Y) _ Cov ()( ⇒ r = (-✓ 312) = - 0.866 (:. byX'bXY <0sor<0) 101(8): 4X + 2y = 1 ⇒ X = - (1/2) )' + (1/4) = -J(-2/3)x(12) . L et r 1- Jlney on x: , y- 85 = 0.487 (x - 36) ⇒ y = 0.48x + 67.72. IO0(A): Let the lines of regression of Y on X and X on Y be x + 2y = 5 and 2x + 3y =8 respectively. Then br.r= - (112) and bxr = - (3/2). r 2 = brx· bxr =(3/4) ⇒ r2 = (3/4) ⇒ b,'X ' correlation coefficient b I be new values of X and y ,,::t"'cch + (116) 0.66x8 X (1 y (1 t ' •11e11 '' I~ =~=---ii-= 0.48. bxy = - (1/2). 2x + 3y = 4 :⇒ )' = - (2/3) Cov (X Y) y , we get : • cz ai Cov (u, v) = (1/n) r (u - ii) (v - v) . Common Proflc1enc 5.31 Test: Quantitative Aptitude (S la!it. Ex= 20, Ly= 45, E xy == 220, L x 2 = 100. Ey2 = 485, n = 5 J22(C): Here = - •c :t (x - x)(y ,re co,· (x, y) (1/11) fl Cov (u, ,,) ac Cov (:c, y) = = Cov (x, y) x~(~ ~.v)i I 5 x 220 - 20 x 45 a <1x X c <1y = p (x,y) yon x · 117(8): Slope of line of regression = b = m1 (say) '> Slope of line of regression x on y I =-=m2 (say) bXl' Let 0 be the angle between them, then b -(11 bw) - T/11 - "'2 = ✓100 123(D): Here L x = 30, Ly= 15, Exy = 94, Ex2= 190 "I: xy- (I: x) (:t Y) ~ 2 'C' II ,1., X -(:t x)2 bxy x (1 I b,y) 5x94-30x15 5 X 190 -(30)2 bxy + b;-,,· e"' tan-• b)'X bXJ' - ] ) ~~- ( bXJ' 470-450 950 - 900 . + bJX I= f (given). Smee the two lines of regress10n pass through (X, Y), therefore. 118(A): Here ⇒ b - X=-=Y 1-a 119(A): Here b1 , = 1 and bx,,= 1 4 (·: ⇒ b X=--=Y. - 1-a ( ·: y = x) :c=(1!4)y+(3l4) Now r= ✓b..n· xbyx = R 10 = 50 = 0.2. I 2 120(A): See Text. 12l(B): Here :I: xy = 140, Ex= 25, l:y = 30, n = 5 =_!_ [300 -(52 X 64) / 11) II 3300-3328 = - 0.23. 121 126(8): The mean x, y 1s the pont of intersection of these lines. Solve the equations to get : x = 5, ji = (1/3). ⇒ ⇒ y = (8/)0) X + (66/10) = (4/5). Sunilarly, = (18/40) )' + (214/48) bxy = (9/20) b,,x X 25 =- C:E x) (t J')} 127(A): 10y =Bx+ 66 5 X 140 - 25 X 30 50 = - 25 (l/11) ⇒ 2• r = ✓byx X b_.,. "" 0 .64 x o.83 -b X 138(B): = o.sl3. 3 b-3b=]__ n==y., 2 ⇒ bxb yx- 'xy 2 So, the 5 l =- y +3 3 is the regression equation of X on y and the equation y = 2t - 4 1s that of = 3X - 5 ~ =3 ⇒ p (X , Y) ⇒ {p {X, Y))· = mJ • (·: :. b y.t :::3 byx and (A): Sy= 4x + 30 5 4 b}.t = -. 5 ~ 20x = 9y + 107 ⇒ X "' r2 9 107 9 20 y + 20 => b'I)' = 20 . =b -')' == 0.36. x b yx == ~x~=~ ✓0.36 = - 0.6. 100 = 0.9 140(B): r2 == b"J' x by,= 1.5 x 0.6 Coefficient of allienation p = ✓I -0.9 == 0.32. 14l(D): Since the two lines of regression therefore, always intersect at ( x. _v ). 2:x + 9y- I79=0. =- 5 20 4x=Sy-137=0 13S(D): In the given data b,., = - 0.!J is negative and b,y = 0.4 is positive which 1s not possible as bn and b,, are always of same sign. Hence the given data is i~consistent. 136(C): {p(X, Y)}2=b<>x by_,=(- 0.4)(-0.9) 2 ⇒ y = ix+ 6 bn. are positive, therefore, p (X, Y) > 132(A): Cov (X, Y), b,,. b,y ,and p (X, Y) all are either positive, or O or negative s1mutaneously. 133(0): See Text. 134(D): See Text. ~. ,'I)' 0) ⇒ p (X, Y) =.!.2 = 0.75. 139 X == = 2 and b~. Y on X. So, bn ··• 3 , 2 2 ⇒ by_, x b.'CJ' y=.!_x +25 2 Coefficient of determination = r2 = bp x b,,, = (1/2) x (3/2) = (3/4) If we take y = 3x - 5 as regression 1Jl(M equation of Y on X and y = 2x _ 4 as that of X on Y, then J X- l( 2 == 1 - r2 = 1 - (0.14)2 = 0.02 . Standard error of estimate of x on y ' 112 j.iO(C)• ~ <1x (1 - r2) = 11 X [I - (0.6)2]1'2 ~ I I x 0.8 = 8.8. not possible, .6 = 0.9 < 1. SO => 2y = X + 50 • _yi SitniJarJy, x = ~ Y - 5 => b = coefficient of non-determination . O b b =15 ~ b J"' f{ere r - XJ' == 0 24 x 0.58 = 0,14. 1l9(JJ): which in equation t :)I I 125(A): Cov (x,y) = (1/11) p: xy - :s1.JJ)= 1 50 "'0.4. 5 X 107 - 15 X 35 5 X 55 - (15) 2 bxy ·-=±- = ± 0.5. r = 0.5 as both b_9 • and byx are positive C ov ( x,y ) 6 :::=4 • 20 124(D): Here Ex= 15, E y = 35, .Exy= 107,Ex2 =55 V = aX + b ⇒ X =aX + b (': X = Y) - -( 4S)l ✓40Q = 10 X iQ "" J. X b."CJ' bp. -1 ~ 200 lfere, P OC. Y) is.negative u b b pt "7 p (X, Y) all ha ve the same sign. 137(1)): Note alterua~t b,. b..,, :S 1. Out of given tives only (D) is such Chat . · >' -;::;, f{icient o fd e t ernunation coe "" ,:i == bxy x b}'X ..1 5 (485) X 1100 - 900 -- .I)' 1 + b1" tan 9 - 1 + m1 m2 = ✓5 (100) - (20) 2 == (6/IOf. :::> (4 /5)=(6/I0)x(cr>'/3) •. r - ~ ) -----...___ [✓ n I: x- - (.t x) 2 p(u,v)-Jvar(u) xJVarM .fi_4 I 5) x (9 I 20) rJoW yx " I: xy - (.t x) (l: - . := b ==(r<J,l<Jx) Solve these two equations to get : x and jias 13 and 17. 142(B): Since p2 $ I, therefore, byx b-'>. :St. 143(A): The two lines of regression are ••• (1) X + 4y = 3 ••• (2) and 3x + y = 15 If we take (I) as the regression equation ofY on X, then (2) is that of X on Y. These two equations can be written as : y =- 1 3 I 15 4 x + 4 and x = - 3Y + 3 common Proficiency Tes : MD JSO(C): 1 __ .!.. ~ b)T = - -4 and bx,· - 3 x,· 12 is valid. To find x, wheny =3, wer are to use the regression t>quation of X on Y. 1 3 144(D): Given data is inconsistent ~s b,.x < 0 and b > o. For any bivariate data both and b should be of same 6p X1 sign or both should be 0. 5(C): If 4x _ 5y + 33 = 0 is the regression 14 equation of yon X, then b,, = Slope = -4 5 1 and the = slope of the other bxy -line = 2 9 ⇒ bxy = 9 36 10 > I. So, our selection is not correct. The regression equation of Y on X is 1x- 9y • 127 = 0 and hence byx = slope = 2/9. I 46(D): Though p2 = b,x x b_ry, therefore, ⇒ Thus, if if IpI= Jbyx bxy b,, >0 b,, <0 is not correct. 147{B): Regression line of Y on X is y- y = byx (x - x) Clearly, slope of this line = byx· 148(A): Solving the lines, we get x = 130, y = 90 as their point of intersection, which is their means (x, y) . x = 130, y = 90. rcry 149(B): b = Y-' cr-' ⇒ -3 (- ✓3) -= 4 4 => cr-' = fj or o/ , ' rines ,_::I 5x 46.'ji,.So Solving then, we get : x "" 1 , ' I I and , y"' I52(A): 4y- 5x = 15 ⇒ Y == (514) X4 'I ' .. b. = 5/4 = 1.25. Also ..., (Is,~ >·• r"" b ') ⇒ (0.75) 2 = bry x 1.25 ~ b -t> ~ b t53(C): The two lines of regressio -9"" n are . 5x + 3y = 55 and 7x + y = 45 ··· (1) If we take hne ( l) as the ··· (?j equation of Y on X, then (i;e&ressi0ii 151 X on Y. We can wnte these as.hat 0r 5 55 . y= --x+- 2 - - x2 crx = 16/3. 3 1 45 = - 7Y +7 -5 =3 =-)jx~=-Ji. 156(B): r2 = b_-ry X byx = 0.64. Coefficient of Determination = ,i =0.64. 157{B): Coefficient determination : r2 = 0.9025 ⇒ r = 0.95. 158{B): r2 = byx X b-'Y = 0.9025 ⇒ 0.39 x b-ry = 0.9025 bxy = 0.9025/0.3 9 = 2.314. 159(B): k2 = 1 - r2 ⇒ 0.502 = I - r1⇒ r2 = 0.498 ⇒ r = 0.11. 160(A): k 2 = 0.502 ⇒ k 161(A): k = 0.4 ⇒ ⇒ = 0.11. k2 = 0.16 1-r2 = 0.16. I"" 1-0.16=0.84 ,-;:.-;;; ,::; 161(C>:, • ✓ 0,84 r"" = 0.9170 (l-,2)1'2 CJ ,-ch ♦ 61'-c>: Sr% ""o.~ (1 - o.16)"2 =0.6 x .Joii. 84 1 ""o.6 x o.917 = o.ss. · · 2 .,, Ox (1-r2)1' 63(JJ): S,cv ""4 x (I - I)= 0. 1 cry ra ,::; , - ⇒ (JJC =--2'.. 6'(1'): bxy 1 <J x - byx == erX (1 - r)112 == o• <Jx b "" r 165(,.\): 11 ⇒ 0. 6 = (0.5) x <J Y -::::1 <J x == :)I- ⇒ cr -..!. OY a,= 3a, = 3 x 4 = 12. 850 (I - r)IIZ = 3.12 X (1- 0.25)112 CJ X 6(/\): Since the two lines of regression meet in (x, y) . therefore, the point (x, ji) always lies on the two regression Jines. (0): The two lines of regression coincide 167 if pz = I, le., if p = - 1 or I. (A) (130, 90) (C) (8~, 120) (B) (90, 130) (D) (120, 80). 2. The equations of two regression lines are as follows : 3X + 2Y - 26 = 0; 6X + Y - 31 = 0. The regression coefficients; (b,y, byx) are: (A) (- 1/6, - 3/2) (B) (1/6, 3/2) (C) (3/2, - 1/6) (D) (-- 3/2, J/6). 3. You are given the following data for the variables x and y. x = 36, y = 85, cr = J I cr = 8 and r = 0.66. The value of X > y x when y = 7 5, is : =-3.3. t 70(D): Given bxy = 0.25 = .!_ <l 4' Also byx bxy⇒ 1 -4 byx <- I ⇒ byx~4. Also, by, and b,y are of same sign, therefore b > 0 i.e., O < b S 4. QUESTION BANK - 6 MULTIPLE CHOICE QUESTION Select the correct alternative out of the given ones : t. For the regression lines : 2y-x - 50 = 0, 3y - 2x - IO = 0, the mean values (x, y) is : _ 30x400 10 196- (J0)2 10 = 3. J2 X 0.87 = 2.7}. 16 ax 169(A): Here,b = Ixy-(txty)/n P Ix2-(tx)2 /n (0.6 x 2.6) I 0.5 = 3.12 :. S,cv == 0/3. Nowfrom(l)b,..=3 ⇒ pa, =3 3 S ,CY => b*•3 Also the - (I) rear-ion Yis x' = 1/3 l equation of X OD :. b = 1/3· F z, , -{l) 0 ~ (1) ~ (2), we get: ,. bxy - I ⇒ p2 = 1 ⇒ p = 1 12 -Jx-=4- respectively. and bxy ::: .:2_ 7 p = • .Jbyx bxy (·: bY.t < 0) Hence, byx ⇒ 3 154(C): When the two• lines are at right anges 1 then there 1s no lmear correl . ' . ation between . _ the two Imes of regres s1on, z.e., p- 0. lSS(C): If p (X, Y) = 1, then the two lines 0I regression coincide. !pl =.Jbyxbxr -{ Jb,, b') p - -Jb,. b,y 2x + 3y = - and x 2• This would mean byx x b_"> = ✓J - p or O~s . x=--x3+5=4. •• CJJ' t51(A): (x, y) lies on the given I < l. So, our choice b xb =r, S,,.. = 2 '" '" (A) 25.93 (B) 26.93 (C) 27.93 (D) 24.93. 4. The following data give the correlation coefficient, means and standard deviation of rainfall and yield of paddy in a certain tract : Annual rainfall Yield per in cm. hectare in kgs. 18.3 Mean 973.5 2.0 38.4 S.D. Coefficient of correlation = 0.58. The most likely yield of paddy when the annual rainfall is 22 cm, other factors being assumed to remain the same, is : (A) 1024.7 (B) 1014.7 (C) 1114.7 (D) 914.7.